HOME ABOUT G.A.C.P's RESEARCH TOOLS JOIN US LINKS CONTACT
Objectives 1 - 4
.
OBJECTIVE ONE
Introduction and Evaluation of New/Improved Lines of Cumin and Milk Thistle.
At present, growers attempting to produce cumin in Saskatchewan typically obtain seed from Syria and/or Turkey which are also the dominant suppliers of this commodity at this time (2008). The material obtained so far from these sources has been too short of stature to be mechanically combined and too late to achieve consistent yields under the relatively short Saskatchewan growing season. Blossom blights have also been problematic in the limited number of cumin lines that have been tested in Saskatchewan.
Although milk thistle has been used medicinally for centuries, it has only been under cultivated production for a few decades. Most of the milk thistle planting material presently available to growers was derived by selection and multiplication of plants from wild stands. Efforts to enhance agronomic perform or quality of milk thistle through systematic breeding have been limited. Consequently the crop has considerable potential for improvement in terms of yields, growth habit, quality (as indicated by silymarin content), uniformity of maturity and particularly in its resistance to shattering.

Screening trials represent a highly cost effective means for identifying material adapted to specific environments, production methods or markets. The project accessed promising lines of both milk thistle and cumin from local sources and areas where these crops have a longer history of production (India, the Middle East and Eastern Europe for cumin and Central Europe for milk thistle). The Plant Gene Resource Centre (PGRC) of Agriculture Canada located in Saskatoon (Dr. A Diederichsen B Curator) assisted by accessing worldwide germplasm collections of these crops.

This project evaluated all available lines of cumin and milk thistle under typical Saskatchewan production conditions – looking for promising lines or parental material for future improvement programs.

CUMIN GERMPLASM TRIALS
links open in a new window/tab
- 2003
- 2004
- 2005
MILK THISTLE GERMPLASM TRIALS
links open in a new window/tab
- 2003
- 2004
- 2005
NEW MILK THISTLE LINES RELEASED BY THE UNIVERSITY OF SASKATCHEWAN
- read more

you have reached the end of objective 1 - Click here to Return to the top of the page

 

.

OBJECTIVE 2:
Pathology Support to Reduce Losses to Disease in Spice Crops.
-pdf

1. Pathogen Isolation from Coriander
Jill Thomson, Department of Plant Sciences, University of Saskatchewan

Although Duczek and others have attempted to identify the organisms responsible for blossom blight of spice crops, there is still some confusion as to which organisms are involved and especially whether the disease may actually represent a complex of organisms that varies with location and prevailing conditions. In 2005 and 2006 one particular fungus was consistently observed on blighted blossoms of coriander at the University test plots. The fungus was isolated from these blossoms in 2005. Cultures of this fungus were tested for pathogenicity in the greenhouse.

Materials and Methods:
Eight pots of mature coriander plants with some flowers present were inoculated with a spore suspension made from 16 day old culture plates of the unidentified fungus. The suspension contained 1.5 x 10 5 spores/ml plus the surfactant Tween 20.  The spore suspension was misted onto the flowering parts of the plants until water droplets were present. The pots were placed in plastic bags to increase relative humidity.  Four pots were kept covered for 48 hours and four pots were covered for 72 hours. Another four pots were sprayed with distilled water as controls, and two of these were placed in plastic bags for 48 hours to increase humidity.

Results and Discussion:
Typical blossom blight infections developed on plants that were inoculated and bagged for both 48 and 72 hours. Several seed heads from the control plants that were bagged, developed aborted fruit, probably because of higher temperatures developing in the bags, but no spores were produced on these fruits. Infected florets were collected from inoculated plants five days after inoculation and were incubated at room temperature and high humidity for three days. Spores typical of the isolate were observed in ooze formed on the florets. Photographs were taken of infected florets removed 11 days after inoculation and incubated in a moist chamber at room temperature for two days. The spores were identical to those used in the original inoculation, indicating that the fungus originally isolated from blighted areas was capable of causing blossom blight symptoms on inoculated plants. Blighted flowers were formed on plants bagged for both time periods, indicating a moist period of 72 hours was sufficient for disease development. The fungus was submitted for identification to the National Fungal Identification Service (NFIS), AAFC, Ottawa. A number of fungi have been implicated in blossom blight of coriander , including AureobasidiumBotrytisFusarium, Alternaria and Sclerotinia. The fungus submitted to Ottawa was identified as aMicrodochium species.  This may be a synonymous with the Aureobasidium species that was originally isolated by Lorne Duzchek. Other Microdochiumspecies, such as M. bolleyi, are known to be synonymous with Aureobasidium species (A. bolleyi, in this case).  However the NFIS wish to confirm this identification with a new isolate, and if possible this will be done in 2008. It is extremely interesting to note that in 2003, Dennis reported a new disease of coriander associated with a Microdochium species, occurring in southern Australia (Dennis, J.I. 2003. New disease of coriander in Australia associated with a Microdochium species. Plant Pathology 52: 408).  It is possible that the Australian and Canadian diseases may be caused by the same pathogen, particularly if the original coriander seed used in the two countries came from the same location. This matter will be discussed with the author of the Australian paper. This fungus was not observed on blight infected caraway blossoms isolated from the same field as the coriander. An Ascochyta species is commonly found on infected caraway blossoms, as reported by Lorne Duzchek.

2. Evaluation of Foliar Fungicides for Control of Blossom Blight in Coriander and Caraway
Jill Thomson and Doug Waterer Summary

Coriander and caraway fields in Saskatchewan are commonly affected by blossom blight resulting in reduced seed yields. Although a single application of the fungicide Quadris (azoxystrobin) is registered for control of blossom blight of coriander, experience in research plots in 2005 and growers’ fields indicate it does not provide adequate control of this disease.  Two applications of the product, at different timings, were evaluated for control of blossom blight of both coriander and caraway in field trials in 2005 and 2006.  Disease levels on coriander were significantly reduced by two applications of Quadris or other fungicides but seed yields were not improved.  Multiple applications of the new fungicide Proline (prothioconazole) provided a high degree of crop protection in the 2007 trial. Disease levels on caraway were not affected by fungicide applications in 2006 or 2007.  Early application of the fungicide appears to be a critical factor for disease control in caraway. A fungus associated with blossom blight of coriander in 2005 and 2006 was isolated and was found to cause disease symptoms characteristic of blossom blight when inoculated on plants under greenhouse conditions. The fungus was identified by the National Fungal Identification Service (NFIS), AAFC, Ottawa, as a Microdochium species. A report of a similar fungus occurring on coriander in Australia was made in 2003.

 Background
Coriander (Coriandrum sativum) and caraway (Carum carvi ) represent economically attractive alternative crops in Saskatchewan (Figures 1& 2). The ten-year average acreage of coriander in Saskatchewan is approximately 17,000 acres, with 22,000 acres grown in 2005.  Caraway production in Saskatchewan ranges from 10,000-20,000 acres, with 12,000 acres of biennial caraway grown in the province in 2005.  Production of both these crops can be severely limited by blossom blight, a disease that attacks the crop during flowering, destroying the flowers and thus reducing seed yield (Figures 1, 2, 3 & 4). Initial studies by Dr. Lorne Duczek (formerly of AAFC, Saskatoon) in 2000 to 2002 revealed that a complex of fungal pathogens is responsible for blossom blight of coriander and caraway.  Although the symptoms of blight are similar, the main causal organisms for blossom blight differ for these two spice crops. In coriander the main fungal pathogen associated with blossom blight was identified as an Aureobasidium species, while in caraway the main pathogen was an Ascochyta species. However both these fungal organisms can be found on either crop, and a number of other fungal pathogens were also found on blight infected blossoms –Fusarium avenaceumF. poaeF. culmorumF. equisetiF. sporotrichioidesF. graminearumBotrytis cinereaSclerotinia sclerotiorum, and Alternariaspecies.  In 2002, Dr. Duczek conducted an initial evaluation of the potential to control blossom blight of coriander and caraway by application of the fungicide Quadris (azoxystrobin).  He concluded that although a single application did reduce disease symptoms, this effect was not maintained to the end of the growing season, and seed yield was not increased by the fungicide application.  A single application of Quadris is now registered for control of blossom blight of coriander.

Evaluation of fungicides for control of blossom blight of coriander and caraway
Initial trials conducted by L. Duczek indicated that the fungicide azoxystrobin (Quadris) can reduce levels of blossom blight in coriander but the effect of a single application of the fungicide did not continue throughout the season.  Further work by D. Waterer (Dept of Plant Sciences, Saskatoon) in 2004 also concluded that a single application of Quadris did not provide effective control of blossom blight in coriander.

Initial studies on control of blossom blight in caraway in 2004 were not successful as the annual caraway used in the trial did not flower until late in the season, no disease developed and it was not possible to harvest the crop before snowfall.

The possibility of controlling blossom blight of coriander by multiple applications of different fungicide combinations, rates and timing was evaluated in field trials at the University of Saskatchewan in 2005, 2006 and 2007.

A stand of biennial caraway was planted in 2005, and was used for evaluation of the effectiveness of Quadris applications for control of blossom blight of caraway in 2006 and 2007.

Coriander 2005 trial Materials and Methods:
A large block of coriander was seeded May 28, 2005 at approx 20 lb/acre, using a seed drill, into land at the University of Saskatchewan that was previously planted to coriander in 2004, 2002 and 2001.  This site was chosen because disease was present in the crop in 2004 and thus it was assumed that disease inoculum would be present in 2005. It was necessary to use a site where infection had occurred previously because it would be difficult to simulate natural infection by application of spore suspensions of the complex of pathogens that apparently cause blossom blight. Crop emergence was noted on June 13th and while the plants were still small, 1 x 5m plots were marked out using a rototiller. Seven plots in six replicate blocks were prepared, to accommodate six fungicide application treatments and one control treatment.  The treatments were arranged in a randomized block design. Overhead irrigation was applied several times during the season, but in general there was sufficient rainfall for crop growth and disease development. The fungicide treatments used on the coriander crop in 2005 were:

  1. Control (no fungicide applied)
  2. Quadris (250g/L azoxystrobin,  Syngenta), applied once at early bloom at 280 g ai/ha
  3. Quadris applied twice (early bloom and mid-bloom) at 280 g ai/ha
  4. Quadris applied three times (early, mid and late bloom) at 90 g ai/ha
  5. Headline EC (250g/L pyraclostrobin, BASF) applied at early and late bloom at 100 g ai/ha, plus Bravo 500 (500 g/L chlorothalonil, Syngenta) applied once at 1250 g ai/L at mid bloom.
  6. Lance (70% boscalid, BASF) applied at early and late bloom at 300 g ai/L, plus Bravo applied at mid bloom.
  7. Dithane (75% mancozeb, Dow AgroSciences) applied at early and late bloom at 1800 g ai/ha, with Bravo applied at mid bloom.

 

The full rate of Quadris used is slightly higher than the highest rate used in canola for control of Sclerotinia stem rot (250 g ai/ha).  On December 19, 2005, the 280 g ai/ha rate was approved via Minor Use Registration for control of blossom blight of coriander.  All the other chemicals were applied at the manufacturer’s recommended rate except for treatment four, when the standard rate was divided into three applications. All chemicals were applied in 400 ml water/18m2 , using a hand-held pressurized application system . The chemicals were applied on July 16 (first flowers opening), August 1 (30-40% bloom), August 14 (seed set started), and the plots were rated for disease on 29 July, 12 August, 26 August , 8 September (11 days – 3 weeks after spraying). The disease was rated on 10 plants selected randomly from each plot using the following rating scheme:

0 – no disease
1 – lesions on leaves, not umbels
2 – lesions on stems, not umbels
3 – 1 umbel on plant affected
4 – >1 but <25% umbels affected/plant
5 – 25% to 75% umbels affected/plant
6 – >75% umbels/plant affected
7 – all umbels/plant affected
8 – plant dead from blight infection.

The average score/plot was calculated and the disease scores analyzed for each assessment date. Blossom blight was observed on volunteer coriander plants close to the plots on July 21st and in the plots on July 29th. The plots were harvested on September using a small plot combine. The seed was cleaned and weighed, and the weight of seed produced per 5 m2 plot was recorded.

Results and Discussion:
Very low levels of blossom blight were recorded at the end of July, but by the second assessment on August 12 significantly more disease developed in the control plots. By the end of August the disease was affecting an average of at least one umbel per plant in the plots where fungicide treatments were controlling the disease and between 25 and 50% of the umbels in the untreated plots (Table 1). By the third disease assessment, on September 8th, more than 25% of the umbels showed symptoms of blossom blight in the fungicide treated plots.  In the check treatments more than 75% of the umbels were infected by this stage of the growing season.  There was significantly less disease in September in the plots sprayed with two applications of Quadris, Headline/Bravo, Lance/Bravo and Dithane/Bravo than in the control plots. The single application of Quadris and three applications of a reduced rate of Quadris did not significantly reduce the level of disease when compared with the control. Despite the high levels of blossom infection, the majority of fungicide treatments did increase the weight of seed produced from the plots (Table 2). Treatments involving two applications of Quadris and the Lance/Bravo treatment more than doubled the plot yield relative to the control. Applications of Headline/Bravo and Dithane/Bravo also significantly increased yield in comparison with the check plots. The least effective treatments were the single application of Quadris and the three applications of Quadris at a low rate.
Table 1.
Effect of fungicide application on blossom blight in coriander on four assessment dates
– Saskatoon 2005.


Fungicide treatment Average disease score on
  July 29 August 12 August 26 September 8
Control .03 a* 1.3 a 5.2 a 6.1 a
Quadris @ full rate X1 0.0 a 0.3 bc 3.6 bc 5.9 ab
Quadris @ full rate
X 2
.07 a 0.8 abc 3.5 c 5.7 bc
Quadris @ 1/3 rate
X 3
.05 a 0.9 ab 4.4 ab 5.9 ab
Headline X 2, Bravo .05 a 0.3 bc 3.6 bc 5.5 c
Lance X 2, Bravo .08 a 0.5 bc 3.2 cd 5.4 c
Dithane, Bravo, Dithane .05 a 0.2 c 2.6 d 5.5 c
Coefficient of Variance 253.8 79.9 19.2 4.8
numbers followed by the same letter are not significantly different, using Duncan's Multiple test,P=0.05

 

Table 2.
Effect of fungicide treatment on seed yield of coriander
– Saskatoon 2005.


Fungicide Treatment Average Seed Yield
Control 0.23 c*
Quadris @ full rate x1 0.38 abc
Quadris @ full rate x2 0.51 a
Quadris @ 1/3 rate X 3 0.33 bc
Headline X 2, Bravo 0.43 ab
Lance X @, Bravo 0.52 a
Dithane, Bravo, Dithane 0.45 ab
Coefficient of Variance 28.6
* numbers followed by the same letter are not significantly different, using Duncan’s Multiple Range test, P=0.05.
Quadris has recently received registration for control of blossom blight of coriander, based on a single application at the early flowering stage. Data from this trial, as well as trials conducted previously by the University of Saskatchewan, suggest that two applications would be needed to provide any significant degree of crop protection. Using only one application it would be extremely difficult to guarantee that the application was made at the correct time, and that disease would be controlled for a sufficient period to safeguard seed yields. The fact that two applications of Quadris (or Lance + Bravo) resulted in a doubling of yield suggests that this level of treatment might be economically justified.  A single application of Quadris costs $20-$30/acre.  Based on this treatment  increasing yields from 400 to 800lbs/acre, with coriander selling at $ 0.30/ lb, spraying twice would yield an extra $120/acre – more than covering the cost of two fungicide applications. Three applications of any product combination did not appear to provide any significant increase in disease control or yield relative to two applications – and therefore would not be justified economically.  The third application came late in the season – at full flowering.  Disease that develops later in the season may not impact yield, as late flowers usually do not produce harvestable seed.  The timing of application in relation to crop stage and the arrival of disease inoculum is critical for economic control. Further trials will examine this relationship to determine the most effective spray program.  For example, a pre-bloom application may be worth considering.
Conclusions:
A reduction in blossom blight levels resulting in significant yield improvements was produced when more than one application of foliar-applied fungicide was made to a coriander crop. A single application of fungicide, or multiple applications at a low rate, did not significantly reduce disease levels or increase yield relative to unsprayed control treatments. Two applications of a number of fungicides may well provide an economically sound control option for growers, but further evaluation of the best timing of application in relation to crop growth and disease onset is necessary.

Coriander 2006 trial
The 2006 trial was conducted at the University of Saskatchewan Horticulture Field Research Station in Saskatoon, on the same plot of land used in the 2005 trial.  This site was chosen because disease was present in the coriander crops in 2004 and 2005 and it was anticipated that disease inoculum levels would be high in 2006. It was necessary to use a site where infection had occurred previously, because it would be difficult to simulate natural infection by application of spore suspensions of the complex of pathogens that cause blossom blight.

A large block of coriander was seeded 1 June 2006, (seed supplied by G. Schweitzer, Eston) at approx 20 lb/acre, using a single cone seed drill. Emergence was noted on June 14th.   Weeds were controlled using a post-emergence application of linuron.  In early July treatment plots (1 x 5 m) were marked out by mowing one meter wide pathways between the plots. Seven treatments with five replicates were arranged in a randomized block design.

There was sufficient rainfall for crop growth and disease development. Irrigation was not applied.

The label rate for Quadris (Syngenta, 250 g /L azoxystrobin) in coriander is 280 g ai/ha.   However, all treatments in this trial used a lower rate (111 g ai/ha) of Quadris – the rate recommended for disease control in pulse crops. This reduced rate was chosen as producers have indicated that it is economically viable to make two applications at this rate.  Previous trials conducted as part of this research project had indicated that multiple fungicide treatments provided superior disease control relative to a single application.  .  The Quadris was applied in the equivalent of 300 L water per hectare. This water volume was higher than would be used commercially but was necessary to ensure adequate coverage using the hand-held pressurized CO2 application type spray system used in this project.

The fungicide treatments used in 2006 were:
1. Quadris (250 g/L azoxystrobin, Syngenta), applied at 111 g ai/ha once at initial bloom
2. Quadris applied at same rate, once at mid bloom
3. Quadris applied at same rate, once at full bloom
4. Quadris applied at same rate, at initial and mid bloom
5. Quadris applied at same rate, at initial and full bloom
6. Quadris applied at same rate, at mid and full bloom
7. Control (no fungicide applied)

The sprays were applied on 25 July when 1% of the plants had flowers petals open (initial bloom), 6 August when 40% of the plants were flowering (mid-bloom) and 10 August when 100% of the plants were flowering and seed set was beginning in the first flowers (full bloom). Disease ratings were conducted on 25 July (before spraying), 4 August, 14 August and 21 August. The disease was rated on 10 plants selected randomly from each plot using the rating scheme described in the preceding section.

The average score/plot was calculated and the disease scores analyzed for each assessment date.

The plots were harvested on 3 October using a small plot combine. The seed was cleaned and weighed, and the weight of seed produced per 5 m2 plot was recorded.

 

Results and Discussion:
Blossom blight was observed on volunteer coriander plants close to the research plots on 25 July, and was seen in the experimental plots on 4 August. At the first disease assessment on 4 August only plots sprayed at initial bloom had received an application of fungicide.  None of these plots differed from the control in levels of disease. On August 14 all the spray treatments had been applied and significant treatment effects on disease levels were observed (Table 3). Both the plots sprayed at either initial or mid-bloom with an additional application at full bloom had significantly less disease than most of the other plots.

 

Fungicide treatment Average disease score on:
  4 August 14 August 24 August
Initial bloom (IB) 0.5 ab* 2.4 a 5.1 a
Mid-bloom (MB) 0.3 b 2.1 a 4.5 ab
Full bloom (FB) 1.2 a 2.7 a 5.0 a
IB and MB 0.9 ab 1.0 bc 4.1 ab

IB and FB

0.1 b 0.6 c 4.3 ab
MB and FB 0.3 b 0.4 c 3.6 b
No fungicide 0.3 b 1.9 ab 4.7 a
Coefficient of Variance (%) 138.4 60.2 19.3

* numbers followed by the same letter in the same column are not significantly different, using Duncan’s Multiple Range test, P=0.1.

At the last assessment the plots receiving two applications of Quadris, at mid and full bloom, still had significantly less disease than the control plots.

None of the fungicide treatments enhanced yields relative to the untreated control.

 

Conclusions:
Two applications of Quadris were more effective than a single application, particularly when one of the applications was when the crop was at full bloom. This supports the findings of the previous year, when two applications of fungicide controlled disease more effectively than a single application. Unfortunately disease control in 2006 did not result in a yield increase, possibly because of the small plot size. It is also possible that the amount of disease controlled was insufficient to impact on yield in 2006, although disease levels between the two years did not appear to differ.
2007 trial
The trial was conducted at the University of Saskatchewan Horticulture Field Research Station in Saskatoon, on the same site as the two previous trials.  This site had been planted to coriander in five of the past six years and it was anticipated that disease inoculum levels would be high in 2007. It was necessary to use a site where infection had occurred previously, because it would be difficult to simulate natural infection by application of spore suspensions of the complex of pathogens that cause blossom blight.

A large block of coriander was seeded 5 June 2007, (seed supplied by G. Schweitzer, Eston) at approx 20 lb/acre, using a bulk seed drill. Emergence of both volunteer and seeded coriander was noted on June 27th.   Weeds were controlled using a post-emergence application of linuron.  In early July treatment plots (1 x 5 m) were marked out by mowing one meter pathways between the plots. Eight treatments with five replicates were arranged in a randomized block design.

There was sufficient rainfall for crop growth and disease development. Overhead irrigation was applied in the evening in early August to further encourage disease development.

The label rate for Quadris (Syngenta, 250 g /L azoxystrobin) in coriander is 280 g ai/ha.   However, all treatments in the 2006 and 2007 trials used a lower rate (111 g ai/ha) of Quadris – this rate corresponds to the rate used for disease control in pulse crops. This reduced rate was chosen as it would probably be economically viable for producers to make two applications at this rate.   The Quadris was applied in the equivalent of 300 L water per hectare. This water volume was higher than would be used commercially but was necessary to ensure adequate coverage using the hand-held pressurized application system used in this project.  Proline is a new product that controls a wide range of pathogens and is registered for use in canola, cereal and pulse crops.

The fungicide treatments used in 2007 were:
1. Quadris applied once at initial bloom
2. Quadris applied once at mid bloom
3. Quadris applied once at full bloom
4. Quadris applied at initial and mid bloom
5. Quadris applied at initial and full bloom
6. Quadris applied at mid and full bloom
7. Proline (prothioconazole, Bayer CropScience Inc., 480 g/L, applied at 400ml
product/ha). Proline was applied at initial, mid and full bloom.
8. Control (no fungicide applied)

The sprays were applied on 29 July when 1% of the plants had flowers petals open (initial bloom), 4 August when 30-50% of the plants were flowering (mid-bloom) and 14 August when 100% of the plants were flowering and seed set was beginning in the first flowers (full bloom). Disease ratings were conducted on 2 August, 14 August and 27 August. There was no assessment made before the first spray date as almost no disease was observed. Approximately 10% of the plants were infected with aster yellows and some Sclerotinia infections were observed. Blossom blight disease was rated on 10 plants selected randomly from each plot using the rating scheme described for 2005:

The average score/plot was calculated and the disease scores analyzed for each assessment date.

The plots were harvested on 4 October using a small plot combine. The seed was cleaned and weighed, and the weight of seed produced per 5 m2 plot was recorded.

Results and Discussion:
Very little blossom blight was observed on either volunteer or seeded coriander plants before the first spray on 29 July. There were no significant differences in the level of disease when the plots were first assessed for disease on August 2. However disease levels had increased by the second assessment date on August 14, 10 days after the second application of fungicides and immediately prior to the third chemical application (Table 3.2.1.4).

There was significantly less disease in the plot that had received 2 applications of Proline, compared with the plots that were either sprayed once with Quadris at early and mid bloom, or not sprayed at all. However, at this date the plots to be sprayed at full bloom had not been sprayed, and the disease levels in these plots were not significantly different from either the plots sprayed earlier with Quadris or the Proline sprayed plots. By the third assessment at the end of August (13 days after the last chemical application) Proline treated plots had significantly less disease than control plots, and so did the plots with one late application of Quadris. It appeared from this trial that the later applications of Quadris were more effective than earlier ones and that Proline was effective when applied three times.

 

Table 4.
Effect of timing of application of Quadris on blossom blight in coriander on three assessment dates – Saskatoon 2007.

 

Fungicide treatment Average disease score on:
  2 August 14 August 27 August
Quadris at initial bloom (IB) 0.1 0.7 a* 2.5 ab
Quadris at mid bloom (MB) 0.2  0.7 a 2.2 ab
Quadris at full bloom (FB) 0.1 0.4 ab 1.9 bc
Quadris at IB and MB 0.1 0.4 ab 2.3 ab

Quadris at IB and FB

0.1 0.4 ab 2.1 b
Quadris at MB and FB 0.1 0.6 ab 2.2 ab
Proline at IB, MB and FB 0.1 0.2 b 1.0 c
No fungicide 0.1 0.7 a 3.1 a
Coefficient of Variance 164.7 66.2 32.7
* numbers followed by the same letter in the same column are not significantly different, using  LSD test, P=0.1.

Timing of fungicide application was obviously important but as it is difficult to be sure of the most effective time to spray.  It was clearly preferable to spray twice to increase the chance of control. Proline obviously has the potential to control blossom blight, but timing was not evaluated in this study. It is important that new chemistries be tested and included in Minor Use registrations as quickly as possible, to ensure producers have effective management options should the pathogens involved develop resistance to the other chemistries.

None of the fungicide treatments enhanced seed yields relative to the untreated control.  This was not surprising as disease scores were generally low, with an average of less than 1 umbel per plant affected.

Conclusions:
In the coriander trial conducted in 2007, three applications of Proline was the most effective treatment in reducing disease. As this treatment would be prohibitively expensive, the optimum timing for one or two applications of this product needs to be examined. Every season is made up of different flowering/ weather interactions and it is extremely difficult to determine a precise recommendation for timing of application of Quadris which is presently registered for the control of blossom blight in coriander.  In 2007 the later applications of Quadris appeared to be more effective, probably as August was wet and conducive to disease development, whereas July was hot and dry and therefore not favorable for the development of blossom blight.

Overall conclusions for fungicidal control of blossom blight of coriander:
Two applications of Quadris, or another fungicide, are necessary for control of blossom blight. However this control of disease does not always translate into an increase of yield, in the small plot trial. Timing of the applications is critical in determining efficacy, and this timing may vary depending on crop stage, weather conditions and inoculum level.
It is recommended that the Minor Use label be altered to include two applications of Quadris on coriander.

Caraway 2005 trial
Initial establishment of the biennial caraway was undertaken. A block of caraway similar in size to that of the coriander was planted in the summer 2005.

2006 trial

Materials and Methods:
The large block of biennial caraway (seed supplied by G. Schweitzer, Eston)seeded in the research plots at the University of Saskatchewan in summer 2005 was used for the 2006 trial. In early June of 2006 treatment plots (1 x 5 m) were marked out by mowing one meter pathways between the plots. Seven treatments with five replicates were arranged in a randomized block design. The caraway plants developed rapidly in the spring and were at full bloom by 13 June when the first fungicide application was made. This time of application is later than was initially planned.  Two rates of Quadris were used – a low rate of 111 g ai/ga and a high rate of 222 g ai/ha.  Three treatments received an initial application at 111 g ai/ha and three treatments received an initial application at 222 g ai/ha.  A second application of either the high or low rates was made on four of the treatments on 24 June, when only 10% of the plants were still flowering and the majority of plants were setting seed. The fungicide treatments were:

  1. Quadris applied at high rate at full bloom (H/FB)
  2. Quadris applied at low rate at full bloom  (L/FB)
  3. Quadris applied at high rate at full bloom (H/FB) and late bloom (H/LB)
  4. Quadris applied at low rate at full bloom  (L/FB) and late bloom (L/LB)
  5. Quadris applied at high rate at full bloom (H/FB) and low rate at late bloom (L/LB)
  6. Quadris applied at low rate at full bloom  (L/FB) and high rate at late bloom (H/LB)
  7. Control, no fungicide applied

Disease ratings were conducted on 7, 15, 22, 29 June and 6 July.  The disease was rated on 10 plants selected randomly from each plot using the same rating scheme previously described for the coriander. The plots were harvested on 26 July using a small plot combine. The seed was cleaned and weighed, and the weight of seed produced per 5 m2 plot was recorded.

Results and Discussion:
There were no significant differences in disease development between the treated plots on June 7 and June 15, which is not surprising as the first spray application was made on June 13 (Table 5). There were some slight differences between treatments noted on 29 June, but none of the spray applications significantly reduced disease. By the beginning of July, all plots were uniformly infected, with the majority of plants having almost 75% of the umbels showing damage by blossom blight.

Table 5. 
Effect of timing and rate of application of Quadris on blossom blight in caraway on five assessment dates – Saskatoon 2006.

Fungicide treatment Average disease score on:
  7 June 15 June 22 June 29 June 6 July
High rate @ full bloom (H/FB) 0.4 2.0 3.3 5.5 ab* 5.6
Low rate @ full bloom (L/FB) 0.8 1.5 3.1 5.1 b 5.6
H/FB + H/LB 0.6 1.4 3.1 5.2 b 5.6
L/FB + L/LB 0.9 1.6 3.8 5.8 a 5.8
H/FB + L/LB 0.8 1.6 2.9 5.2 b 5.6
L/FB + H/LB 0.7 2.0 3.3 5.3 5.6
No fungicide 0.4 1.8 3.6 5.5 ab 5.6
Coefficient of Variance (%) 64.2 35.7 23.1 8.2 6.3

* numbers followed by the same letter in the same column are not significantly different, using Duncan’s Multiple Range test, P=0.1.

 

Conclusions:
The lack of disease control in this trial is probably due to the late application of fungicides. Typically it is recommended that the first application of fungicides for the control of blossom blight be made by early bloom, and certainly no later than 50% bloom. This trial reinforces the recommendation for early spraying, as late applications did not prevent disease development.

2007 trial

Materials and Methods:
The block of caraway seeded in 2005 and used for the fungicide trial in 2006 was combined in 2006 and allowed to re-grow in 2007. Commercial plantings of caraway are commonly harvested in the second and third year after planting. In early June of 2007 treatment plots (1 x 5 m) were marked out by mowing one meter wide pathways between the plots. Uniform disease development had occurred in the plot in 2006 and disease levels had not been significantly affected by the fungicide treatments tested in 2006. It was therefore assumed that the inoculum over-wintering in the block was uniformly present and that there would not be a carry-over effect from disease control in 2006.  Seven treatments with five replicates were arranged in a randomized block design. The first fungicide application was made on 9 June when plants were already 30% flowering.  The flowering period was short and the third fungicide application was made at seed set, when few blossoms were present.

The fungicide treatments were:

  1. Quadris applied at mid bloom (30-50% plants with first flower open, 9 June)
  2. Quadris applied at late bloom (seed set starting on 50% of plants, 21 June)
  3. Quadris applied at seed set  ( 20% of plants still flowering, remainder set seed, 1 July)
  4. Quadris applied at mid and late bloom
  5. Quadris applied at mid bloom and seed set
  6. Quadris applied at late bloom and seed set
  7. Control, no fungicide applied

Blossoms with suspected blossom blight were sampled on June 28 within the plots, and were incubated under moist conditions at room temperature for 5 days.Ascochyta pycnidia and spores formed on the flower parts, confirming the presence of the blight pathogen.  Infected seed heads were also collected from a separate block of caraway in the same field on July 9 and Ascochyta pycnidia and spores were observed after 24 hours of incubation. This confirms the presence of inoculum in the general plot area.

Plot disease ratings were conducted on 8, 16 and 29 June The disease was rated on 10 plants selected randomly from each plot using the same rating scheme previously described for coriander. The plots were harvested on 8 August using a small plot combine. The seed was cleaned and weighed, and the weight of seed produced per 5 m2 plot was recorded.

Results and Discussion:

Variations in disease level between plots was observed in the first assessment, the day before the first spray application, but there were no significant differences in disease levels between treatments after the plots were sprayed.  The initial spray should have been applied slightly earlier and the following two sprays should have been at weekly intervals. The plants flowered rapidly and the spray window was relatively short.

Disease levels were low in 2007, which was surprising as significant levels of disease had developed in this block of caraway in 2006. It was expected that the disease would carry over in the lower portion of the crop, and that levels of blossom blight would be greater than in the previous year, but this did not occur, despite adequate rainfall.

Fungicide treatments did not have a significant effect on yield.  This is not surprising as there were no observed differences in disease levels.

Results and Discussion:
Variations in disease level between plots was observed in the first assessment, the day before the first spray application, but there were no significant differences in disease levels between treatments after the plots were sprayed.  The initial spray should have been applied slightly earlier and the following two sprays should have been at weekly intervals. The plants flowered rapidly and the spray window was relatively short.

Disease levels were low in 2007, which was surprising as significant levels of disease had developed in this block of caraway in 2006. It was expected that the disease would carry over in the lower portion of the crop, and that levels of blossom blight would be greater than in the previous year, but this did not occur, despite adequate rainfall.

Fungicide treatments did not have a significant effect on yield.  This is not surprising as there were no observed differences in disease levels.

Conclusions:
Fungicide applications had no effect on disease levels in the caraway plots in 2007. Disease levels were low and the second applications were made later than desirable, when seed was already setting. Fungicides should be applied early in June, when the crop starts flowering.  Disease levels should also be tested at weekly intervals, as the total flowering period of caraway may be as short as three weeks.

3. Effect of seed treatments on the germination and establishment of cumin and other aromatic spice crops

Jill Thomson and Doug Waterer

Summary
Problems with stand establishment followed by progressive die back of the crop as the season progresses have limited grower interest in cumin (Cuminum cyminum)as a crop in Saskatchewan.  In 2005 cumin stand establishment in the field was improved when seed was treated with the fungicide Apron Maxx (fludioxonil + metalaxyl). The recommended seed treatment, Maxim (fludioxonil), was not effective in this trial. The effect of five seed treatments and one in-furrow treatment on stand establishment of cumin was evaluated in field trials at two research sites and in the greenhouse in 2006. No improvement in seedling establishment due to seed treatment was observed at either field site. Under greenhouse conditions, seed treatments containing metalaxyl (Apron Maxx and L 1269) significantly increased seedling survival in non-sterilized field soils. When field soils were sterilized by autoclaving an excellent stand was achieved irrespective of the seed treatment applied. In a greenhouse trial in 2007 an interaction was observed between soil type, seed treatment efficacy and sterilization of soils. These results indicated that pathogenic fungi may play a significant role in reducing cumin stand establishment in field soil, but there is a complex interaction of other factors that may impact the survival of cumin seedlings under field conditions.

Introduction:

The establishment of a cumin crop sufficiently dense to provide a reasonable yield has been problematic for both researchers and producers in Saskatchewan. Poor plant establishment may reflect a number of factors including low seed viability, lack of tolerance of germinating seedlings to spring temperature and moisture conditions and vulnerability of seed and seedlings to attack by soil and seed-borne pathogens. The possibility that seed treatment of cumin may reduce pathogen attack and hence increase plant establishment was examined in field trials and greenhouse experiments in 2005, 2006 and 2007.

Field Trials

2005 trial

The cumin seed used in the field trial was obtained from G. Schweitzer (G.H. Schweitzer Enterprises Ltd, Eston, Saskatchewan). It originated in Turkey and had a germination rate of 86% in petri dishes held at room temperature.  This line has shown reasonable agronomic performance in previous trials and can therefore be considered as “adapted” to Saskatchewan conditions.  The seed was divided into 300 g lots, with each lot receiving a different seed treatment product.  Each seed lot was placed in a plastic tub, the seed treatment was added and the tub was then shaken until the seed was evenly coated with product. All chemicals were added at the recommended rates.  The total volume of liquid applied per 300 g seed lot was made up to 15 ml with water. This is a higher water volume than would be used commercially but it was necessary to ensure equal distribution of the product on the seed.  The treated seed was poured into a plastic tray, dried in a fume hood for 18 hours and stored in a plastic container until seeding 2-4 weeks later. Sufficient seed was treated for use in all field trials and greenhouse experiments.

Seven seed treatments were tested.  An eighth seed lot was treated with water only, this served as a check. The seed treatments were:

  1. L1269 (Bayer CropScience), a.i. 10.8 g/L metalaxyl, 13.5 g/L trifloxystrobin, applied at a rate of 7.4 ml product/kg seed

2.  Allegiance FL (Bayer CropScience), a.i 317 g metalaxyl/L, applied at 0.32 ml           product/kg seed

  1. Gaucho 480 FL (Bayer CropScience), a.i. 480 g/L imidacloprid, applied at 45 ml product/kg seed, plus the fungicide Jazz (at rate in treatment #1).
  2. Gemini (BASF), a.i. 1.25% triticonazole, 12.5% thiram, applied at 3.6 ml product/kg seed
  3. Vitaflo 280 (Bayer CropScience), a.i. 15.59% carbithiin, 13.25% thiram, applied at 3.3 ml product/kg seed
  4. Maxim 480 FS (Syngenta), a.i. 40.3% fludioxonil, applied at 2.6 ml product/25 kg seed
  5. Apron Maxx RTA (Syngenta), a.i. 0.73% fludioxonil, 1.10% metalaxyl-M, applied at 9.6 ml product/kg seed
  6. Control, seed treated with 15 ml water.

Treatment with the insecticide Gaucho was included to determine if insects were reducing seedling viability. A seed coating material, (Precise) and talc was used with the Gaucho application and this treatment was combined with L1269 to provide control of fungal pathogens.

Seed treatments of caraway, coriander and cumin were evaluated at the Horticulture Research plots, University of Saskatchewan, Saskatoon.  Seed treatments of cumin were also evaluated at the Potato Research plots in Saskatoon. A similar cumin trial, using four of the seed treatments was conducted at the AAFC Scott Research Farm.

The Horticulture Field Station site features a Sutherland series clay loam soil (pH 7.4, E.C. <1.0 dS/m).  The test site has been used for the production of various spice crops for the past 4 years.  Problems with stand establishment have been noted in previous trials conducted at this site.  The Horticulture Field Station trials for each of the three spice crops were planted on June 8th, 2005, in individual, randomized blocks with four replicates. Two 1.5m rows of each treatment were planted 25 cm apart, with no pathways between replicates. Seed was planted at 10 mm depth into a moist seed-bed, using a disk seeder designed for planting vegetable seed. Plots received irrigation several times during the season and did not suffer from moisture stress. Plots were hand-weeded when necessary. Heavy rainfall in June and July caused extensive wash-out of the cumin plots at this site.  Seedling establishment data was collected from 3 of the 4 replicates on July 6th.  No further data was collected from the cumin plots at this site as the plant stands suffered severe decline early in the season.

The Potato Research Field site features an Asquith series sandy loam soil (pH 7.6, E.C. 1.1 dS/m).  The field was in long-term forage production until 2002 and then was planted to plough-down crops of canola in 2003 and 2004.  Spice crops have never been grown at this site.

The trial was planted on June 6th, 2005, in a randomized block design with five replicates.
Three meter long rows of each treatment were planted 25 cm apart, with 1 meter pathways between blocks. Seed was planted at 10mm depth into a moist seed-bed, using a disk seeder designed for planting vegetable seed. The trial was located at the edge of the potato plots and received irrigation when the potato plots were irrigated. Plots were hand-weeded several times early in the season.

Emergence counts were made for the first 1m of row on July 12, 29 and August 26. Final stand counts were conducted based on 2m of row in the third week of September. This gave the most accurate count of the number of plants present, as in mid-season it was sometimes difficult to tell how many plants were present if branching occurred below the soil surface.

The trial at the Scott Research Farm was planted on May 19th, directly into wheat stubble. Standard fertilizer and weed control practices were carried out. Treatments were planted in 5m long rows in a randomized block design with four replicates. Plant counts were taken in two randomly chosen 1m sections of row, towards the front and back of each row, on June 16th, August 10th and 31st. This data was used to calculate the average plant count/m2. The crop was combined on August 31st and the average yield (kg/ha) was calculated.

Results and Discussion:
L1269 and Allegiance treatments both increased the initial plant count for cumin at the Horticulture Field Station site.

The data presented are for plants/m, as these were the values taken.  The stand density on a per meter squared basis can be obtained by multiplying the values by four. Thus the average plant stand for cumin was 76 plants/m2.
A continuous stand decline was seen in the cumin plots but  not observed in the other spice crops, suggesting that cumin may be more susceptible to soil pathogens, may have more seed-borne problems, and/or may be less adapted to the growing conditions encountered in Saskatchewan.

Table 1.
Effect of seed treatment on initial stand establishment of caraway, 

coriander and cumin – Horticulture Research Station – July 2005.

Seed treatment Average plant count (#/m) on July 6 for
  Annual caraway Biennial caraway Coriander Cumin
L1269 16.8 a* 77 9 ab 118.2 a 27.5 a
Allegiance FL 20.1 a 109.1 a 83.5 bc 29.7 a
Gaucho 480 FL + L1269 17.1 a 73.7 ab 113.9 ab 21.7 ab
Gemini 21.5 a 69.5 b 96.0 abc 10.0 b
Vitaflo 280 20.1 a 95.5 ab 109.5 ab 18.2 ab
Maxim 480 FS 22.1 a 99.6 ab 114.0 ab 12.2 b
Apron Maxx RTA 18.3 a 67.4 b 106.4 ab 18.8 ab
Control (no chemical) 19.6 a 100.3 ab 71.6 c 13.3 b
Coefficient of variance 21.1 27.0 20.0 32.7
* numbers followed by the same letter are not significantly different, using Duncan’s Multiple Range test, P=0.05.

 

Cumin stand counts at the Potato Research site were highly variable, as shown by the high values for the coefficients of variance (Table 2).  This suggests that more replication or bigger plot sizes are needed when studying problems of this type.  The first cumin plant count taken on July 12, approximately 5 weeks after seeding, showed that there were significantly more plants established per meter of row relative to the non-treated control when the seed was treated with either Apron Maxx or Gaucho + L1269 (Table 2). This result is quite different from the result obtained at the Horticulture site. Soil type and conditions differ at these two sites, and the pathogens present, and their response to the seed treatments is likely to vary.

Table 2.
Effect of seed treatment of cumin plant count at the Potato Research  site – July and August 2005

Seed Treatment

Plant count per meter of row

  July 12 July 29 August 26
L1269 23.8 ab* 20.0 abc 16.0 ab
Allegiance FL 24.2 ab 20.0 abc 15.6 ab
Gaucho 480 FL + L1269 32.2 25.0 18.4
Gemini 15.4 b 12.8 c 9.8 bc
Vitaflo 280 15.2 b 14.0 bc 10.6 abc
Maxim 480 FS 12.6 b 11.8 c 7.0 c
Apron Maxx RTA 33.2 a 22.6 ab 18.4 a
Control (no chemical) 15.8 b 11.4 c 9.6 bc
Coefficient of variance 37.3 37.9 43.8
* values followed by the same letter in the same column are not significantly different using Duncan’s Multiple Range test, P = 0.05.

 

The seed treated with Maxim, Vitaflo and Gemini did not establish a better initial stand than untreated seed. Although the seed treated with L1269 and Allegiance did show an increase in establishment, this was not significantly greater than the untreated check. At the Horticulture Field Research site the L1269 and Allegiance treatments had increased plant stand. Further evaluation of these treatments is recommended.

The number of cumin plants per meter declined in all treatments as the season progressed.  Dead and dying plants were observed throughout the season.  Root systems of the dying plants were necrotic and lacked side roots, but it was not clear if this damage was a cause or a result of the top die-off. Isolations from the roots of dying cumin plants at other locations showed the presence of Pythium and Fusarium species but the pathogenicity of these isolates was not tested.  Throughout the season the cumin stand counts remained higher in the rows planted with seed treated with Apron Maxx and Gaucho + L1269 (Table 2). On August 19 it was noted that the plants were setting seed and symptoms of blossom blight were also observed in the trial.

The ranking of the treatments remained the same when plant number per 2 metre of row was evaluated in September (Table 3). Seed yields from this trial were extremely poor, with very small, withered, unmarketable seed being produced. The relative yields reflected the final number of plants per row (Table 3).

Table 3.
Effect of seed treatment on the final plant count and seed yield of
cumin at the Potato Research plot site – September 2005.

Seed Treatment Plant count per 2m of row Seed yield from 2m of row (g)
L1269 43.6 ab* 26.8 ab
Allegiance FL 42.8 ab 25.7 ab
Gaucho 480 FL + L1269 53.8 a 35.1 a
Gemini 24.4 b 15.4 b
Vitaflo 280 25.0 b 17.6 b
Maxim 480 FS 24.8 b 14.8 b
Apron Maxx RTA 49.8 a 24.3 ab
Control (no chemical) 26.4 b 17.3 b
Coefficient of variance 36.6 38.9
* values followed by the same letter in the same column are not significantly different using Duncan’s Multiple Range test, P= 0.05.

 

The increased plant establishment obtained when Apron Maxx was applied to the cumin seed indicated that both metalaxyl and fludioxonil may both be important in increasing plant survival. When Maxim, containing fludioxonil alone was applied, plant survival was not increased relative to the control; however, Maxim was applied at half the rate of fludioxonil present in Apron Maxx.  Allegiance, which was applied at a rate of metalaxyl similar to that in Apron Maxx, did not improve plant establishment as significantly as Apron Maxx. Thus it would appear that the combination of metalaxyl and fludioxonil was most effective in increasing stand establishment. L1269 also contained metalaxyl, in combination with trifloxystrobin.  Allegiance and L1269 treatments increased plant establishment, but not as significantly as Apron Maxx and Gaucho + L1269. The presence of the insecticide in Gaucho appeared to increase the plant establishment when compared with the L1269 treatment alone, indicating that insects may influence seedling mortality. This was somewhat unexpected as there was no evidence of insect damage to any of the plants in this trial.  Gaucho is primarily used to control flea beetles and there was no sign of flea beetle damage to the cumin plants. It is possible that wireworms could be causing root damage, making the seedlings more susceptible to subsequent attack by fungal pathogens, but wireworms are not considered to be a problem at this site. Gaucho was applied with a seed conditioner (Precise) and dried with a talc application, all of which could influence the efficacy of this treatment. The effect of Gaucho with and without a fungicide should be evaluated in future trilas. The chemicals present in Gemini and Vitaflo (triticonazole, thiram and carbithiin) did not improve plant establishment when compared with the check.

At the Scott Farm site, Apron Maxx increased plant count at the June evaluation, while the Maxim treatment had fewer plants emerged than in the control (Table 4).

Table 4.
Effect of seed treatments on cumin stand and seed yield at Scott
Research Farm – 2005

Seed treatment Average plant count/m2 on Average yield 
  16 June 10 August 31 August  
Gemini 110.7 bc* 99.7 a 61.0 a 15.7 a
Vitaflo 280 145.2 b 101.7 a 65.6 a 13.3 a
Maxim 480 FS 102.4 c 70.2 a 45.9 a 10.0 a
Apron Maxx RTA 199.8 a 130.6 a 93.2 a 20.3 a
Control (no chemical) 148.6 b 108.3 a 64.3 a 17.7 a
Coefficient of variance 17.7 47.3 59.8 57.4
* values followed by the same letter in the same column are not significantly different using Duncan’s Multiple Range test, P= 0.05.

 

Stand count declined in all treatments through the growing season at the Scott site, with an average of 76% of emerged plants surviving by August 10th and only an average of 45% surviving through to harvest.  By August there were no significant effects of any of the seed treatments on plant stand, although the Apron Maxx plots always had the highest counts and the Maxim plots the lowest. Very low plant counts were recorded in one replicate of this trial and only the three “normal” replicates were evaluated after the June count. The decrease in replication and increase in coefficient of variance likely explains the lack of statistical significance of the increased counts and yield data for Apron Maxx in August.  Seed yields reflected the final plant density.  The weight of seed harvested from this trial was less than the amount of seed required to sow this trial.

Conclusions:
Stand establishment and subsequent seedling survival were increased when certain fungicidal seed treatments were applied to cumin seed, indicating that pathogenic fungi were responsible for some of the plant mortality observed in cumin stands. However plants continued to die throughout the growing season and the seed treatments did not prevent this further decline in stand counts. This was not surprising as seed treatments do not generally protect plants from pathogens that continue to attack plants as the growing season progresses.

The product currently registered for use as a seed treatment on spices, Maxim 480 FS (a.i. fludioxonil), applied at the recommended rate did not enhance stand establishment in any of the 2005 trials. Apron Maxx, containing both fludioxonil and metalaxyl, was effective in increasing stand establishment at two of the sites (Potato Research site and initially at Scott), suggesting that both chemicals may be necessary for effective control of the pathogens that are damaging the cumin stand.  Metalaxyl is effective in controlling the pathogenic species Pythium and Phytophthora, and fludioxonil is effective against a range of pathogens including Fusarium and RhizoctoniaPythium and Fusarium species have both been isolated from diseased cumin roots and Rhizoctonia is a common soil-borne pathogen of many crops (including canola, pulses and potato). The increase in cumin stand establishment observed when seed treatments capable of controlling these pathogens are used, suggests that all three pathogenic species may be problematic for the cumin crop. Further testing of seed treatment products at different locations is recommended, as pathogen populations are likely to vary between locations, and seasons.

2006 trial

Materials and Methods:
The cumin seed source used in the 2005 trial was also used in 2006.  The seed was divided into 300 g lots, with each lot receiving a different seed treatment product.  Each seed lot was placed in a plastic tub, the seed treatment was added and the tub was then shaken until the seed was evenly coated with product. All chemicals were added at the recommended rates.  The total volume of liquid applied per 300 g seed lot was made up to 10 ml with water. This is a greater volume than would be used commercially but it was necessary to ensure equal distribution of the product on the seed.  The treated seed was poured into a plastic tray, dried in a fume hood for 18 hours and stored in a plastic container until seeding 1-2 weeks later. Sufficient seed was treated for use in both field and greenhouse trials.

Five seed treatments were tested.  An in-furrow application of Quadris at seeding was also evaluated. One seed lot was not treated, this served as a check. The treatments were:

  1. Apron Maxx RTA (Syngenta), a.i. 0.73% fludioxonil, 1.10% metalaxyl-M,

applied at 9.6 ml product/kg seed

  1. L1269 (Bayer CropScience), a.i. 10.8 g/L metalaxyl, 13.5 g/L trifloxystrobin, applied at a rate of 7.4 ml product/kg seed
  2. Gaucho 480 FL (Bayer CropScience), a.i. 480 g/L imidacloprid, applied at 45 ml product/kg seed
  3. Gaucho 480 FL (Bayer CropScience), a.i. 480 g/L imidacloprid, applied at 45 ml product/kg seed, plus the fungicide L1269 (at rate in treatment #2).
  4. Maxim 480 FS (Syngenta), a.i. 40.3% fludioxonil, applied at 0.21 ml product/ kg seed
  5. Quadris (Syngenta), a.i. 250 g/L azoxystrobin, applied in furrow at 6 ml product in 1500 ml water/100m row
  6. Control, seed treated with 10 ml water.

The treatment that combined the insecticide Gaucho with the fungicide L1269 was effective in 2005, so the insecticide and fungicide were evaluated singly and in combination in 2006. Gaucho was originally included to determine if insects were reducing seedling viability. A seed coating material (Precise) and talc was used with the Gaucho application.

Seed treatments were evaluated at the Potato Research plots, University of Saskatchewan and also at the edge of a barley field at Innovation Place, Saskatoon.   Both sites feature a sandy loam soil.  Neither site had been planted to cumin in the past.  Two 2 m rows of each treatment were planted 30 cm apart, with a 2 m pathway between the 5 replicates, in a randomized block design. The Potato Research Plot trial was planted on May 20 at 10 mm depth into a moist seed-bed, using a disk seeder designed for planting vegetable seed. The seeding rate was 100 seeds per meter. The in-furrow application of Quadris was applied into a 10 cm wide by 5 cm deep furrow using a pressurized sprayer. The furrow was filled in with soil and the cumin was seeded into the treated row.

The Potato Plots received irrigation several times during the season and the cumin did not suffer from moisture stress. Plots were hand-weeded when necessary. The Innovation Place site was planted on May 27 and did not receive irrigation. Weed pressure was high at this site.  Despite hand weeding, the combination of weed competition and moisture stress reduced the vigor of the cumin plants at this site.

Plant counts were made on July 5, 19 and August 4 at the Potato Plot site. Final stand counts were conducted when the plants were removed on September 1. This gave the most accurate stand count, as in mid-season it was sometimes difficult to tell how many plants were present if branching occurred below the soil surface. Plant counts were made on July 18, August 2 and 16 at the Innovation Place site. The numbers were so low and the plants so small by the August 16th count that no further counts were taken.

Results and Discussion:

Field trials:
There were no significant differences between the average plant counts for any of the treatments at all four sampling times (Table 3.3.1.6) at the Potato Plot site. However the Gaucho plus fungicide treatment had the highest score at all times.

It is interesting to note that there was a decrease in plant numbers from early July to mid July, but plant counts remained stable from then on. Even at its peak, the stand was relatively poor (average of 19/m) – considering that the crop was seeded at 100 seeds/m.  When the plants were removed and counted at the beginning of September there appeared to be a slight increase in plant numbers compared with the counts in August. This apparent discrepancy between the counts was probably due to the fact that a cluster of individual plants closely resembled a single robust plant, and was likely counted as such in the August counts. The more accurate count was obtained when the plants were removed and counted.

The stand data from the Potato Research site was highly variable from replicate to replicate, as shown by the high values for the coefficients of variance (Table 5).  Two rows were planted for each treatment, and five replicate blocks were used, in an attempt to manage this variability.

Table 5. 
Effect of seed treatments on cumin stands – Potato Research Site, 

2006.

Seed treatment Average plant count (#/m)
  5 July 19 July 4 August 1 September
Apron Maxx RTA 15.0 9.8 10.1 12.1
L1269 16.5 9.9 11.4 11.1
Gaucho 480 FL 20.6 14.3 11.1 16.3
Gaucho 480 FL + L1269 21.0 16.2 17.6 20.2
Maxim 480 FS 19.4 15.5 15.6 16.4
Quadris in- furrow 19.0 13.7 13.1 16.1
Control (no chemical) 20.2 9.4 10.7 11.7
Coefficient of variance (%) 39.6 50.9 49.4 49.0

In 2005 the presence of the insecticide in Gaucho appeared to improve stand establishment when compared with the L1269 treatment alone, indicating that insects may influence seedling mortality. In 2006, the highest plant counts were again obtained with the Gaucho plus fungicide treatment. As mentioned in the introduction, this was somewhat unexpected as there was no evidence of insect damage to any of the plants in this trial.  Flea beetles were observed in the plots and Gaucho is primarily used to control flea beetles.

Yield data were not taken in this trial, as in 2005 seed yields were found to reflect the final plant density. The plant stands were so poor in the 2006 trial that very little seed would have been obtained.

Stand establishment was even more limited at the Innovation Place site than at the Potato Research site – less than 10% of the seeds planted had emerged by July 18 and the stand declined over the remainder of the season (Table 6).  There was no advantage to treating the seed at this site – in fact, the treated seed did not establish as well as the untreated seed. Quadris applied in-furrow was the only treatment that did not reduce plant counts relative to the control.

Table 6.
Effect of seed treatments on initial and mid-season stand 
establishment of cumin, Innovation Place Site, 2006

Seed Treatment

Average plant count (#/m) on:

  July 18 August 2 August 16
Apron Maxx RTA 6.2 b* 4.7 bc 3.9 bc
L1269 6.1 b 3.9 cd 3.1 cd
Gaucho 480 FL 4.8 b 2.7 d 2.0 d
Gaucho 480 FL + L1269 5.4 b 3.2 cd 2.8 cd
Maxim 480 FS 5.4 b 4.0 cd 3.7 bc
Quadris in- furrow 8.3 a 6.0 ab 5.8 a
Control (no chemical) 9.2 a 6.4 a 5.1 ab
Coefficient of variance (%) 25.7 28.7 34.9
* values within a column followed by the same letter are not significantly different using Duncan’s Multiple Range test, P = 0.1

 

 

 

 

 

 

Conclusions:
Overall stand counts were low at both sites and crop establishment was very poor. None of the seed treatments increased stand establishment significantly, unlike the findings in 2005. These results demonstrate the importance of testing treatments at different sites, over multiple years.

2007 trial

In trials conducted in 2005 and 2006, there had been only marginal responses to fungicidal seed treatments applied to cumin.  Despite the use of “good” quality seed, stands had been poor from the onset and had further declined over the course of the growing season.  One possible explanation for the poor quality of the initial stand was the use of seeding equipment in 2005 and 2006 that was best suited to small plot trials.   It was decided to utilize larger scale seeding equipment in 2007.

Materials and Methods:
The cumin seed used in the previous trials was used again in 2007. Only one seed treatment, Apron Maxx, was used in an attempt to obtain a reasonable stand.  One kg of seed was treated with Apron Maxx, at 9.6 ml product/kg in 50 ml water. The seed was treated in a plastic bag, shaken to ensure even coating, and allowed to dry for several hours before planting. Seed was planted using a six row seeder. The seed was placed in the two fertilizer boxes, one box with treated seed and the other with untreated seed, so that six rows were planted in one pass, with three rows of treated seed beside three rows of untreated seed. A block of 12 plots, in a three by four design, was planted, with a 6 row X 3m pass made in each plot. The plots were separated by 1m pathways. This trial design was used so that if adequate stand establishment occurred the plots could then be used for a fungicide trial, to evaluate chemical control of blossom blight.  The seed was planted to a depth of 2.5 cm to reach moisture – this may have been deeper than optimal for cumin.

Results and Discussion:
Cumin emergence was noted three weeks after planting. The number of plants/3m, in each of the 6 rows planted per plot, were counted 5 weeks after planting. When the average seedling establishment for the rows was compared there was no significant difference between the untreated and treated seed.

Seedling establishment was generally low and quite variable from replicate to replicate. The plants continued to die during the summer and there were inadequate plant numbers to use in the proposed trial for control of blossom blight.  Until an adequate stand of cumin can be obtained it was not possible to evaluate other diseases in the crop.

Conclusions:
The use of larger seeding equipment did not result in a better stand establishment, and the seed treatment with Apron Maxx did not improve stand establishment at this site.

Overall conclusions for efficacy of seed treatments on cumin stand establishment in the field:

The field trial in 2005 indicated that cumin stand establishment was improved the most when seed was treated with Apron Maxx. The recommended seed treatment, Maxim, was not effective in the 2005 trial. However seed treatments were not effective in the two following years, indicating that poor stand establishment may be caused by a variety of factors that vary between sites, seasons and other conditions. Further work is necessary in order to gain an understanding of the factors affecting stand establishment. Manipulation of these factors may then lead to the establishment of satisfactory stands of cumin.
Greenhouse trials:

2005 trial

Cumin, coriander and caraway plants were grown in the greenhouse in a pathogen-free soil mix to determine if problems in stand establishment were occurring because of pathogens in the soil or on the seed.  If the pathogens were present on the seed then disease would be most severe on plants that had not been treated with a protectant fungicide applied to the seed.

Materials and Methods:
Samples of the seed treatments tested in the 2005 field trials (reported previously) were also evaluated in the greenhouse. Ten 4” pots were planted with 10 seeds from each seed treatment.  The pots were filled with potting soil (Sunshine mix #3).   Although this mix is not sterile, it harbors very few pathogenic organisms.  The pots were arranged in a randomized block design in a greenhouse maintained at 24°C/18°C day/night.  The pots were watered as necessary.  The number of plants emerged per pot was recorded 17 days after planting.  The pots planted with cumin were maintained in the greenhouse for a further 12 weeks to determine if the stand decline observed in field trials also occurred under greenhouse conditions.

Results and Discussion:
The germination percentage for the cumin under the greenhouse conditions used in this trial was far higher than that observed in the field trial (Table 1).  This suggests that;

  • field conditions in the spring were not suited to this crop or,
  • pathogens present in the field soil were killing the seedlings prior to emergence.

Table 1.
Effect of seed treatment on initial plant count of spices in the greenhouse – July 2005.

Seed treatment

Average germination of 10 seeds 22 days after planting:
  Coriander Annual caraway Biennial caraway
L1269 8.8 a* 3.0 ab 6.4 ab
Allegiance FL 9.9 a 3.7 a 5.8 ab
Gaucho 480 FL + L1269 8.0 a 3.2 ab 5.2 b
Gemini 8.6 a 2.9 ab 6.9 a
Vitaflo 280 8.1 a 1.9 b 5.6 ab
Maxim 480 FS 8.4 a 3.1 ab 5.8 ab
Apron Maxx RTA 8.3 a 3.0 ab 5.8 ab
Control (no chemical) 9.4 a 2.7 ab 6.7 ab
Coefficient of variance 22.2 53.8 25.3
* values followed by the same letter in the same column are not significantly different using Duncan’s Multiple Range test, P= 0.05.

Germination of coriander was also excellent, while the annual caraway had a low germination %, irrespective of the seed treatment used.

The seed treatments had no effect on germination of coriander planted into “disease-free” soil-less media (Table 1).   As seed treatment effects were observed in the corresponding field trials, this suggests that the cause of some stand establishment problems seen in coriander is pathogens present in the soil. In the annual caraway, the Allegiance treatment increased plant counts in comparison with the Vitaflo treatment, but none of the treatments significantly improved plant counts relative to the untreated control.  A similar result was observed in the biennial caraway and the cumin pots. There were no significant treatment effects for the final plant count for cumin after 12 weeks (Table 2).  It is noteworthy that no decline in plant counts was observed over the 12 weeks that the cumin grew in this trial.  Over 50% of the plants died during a corresponding period of time in the field.  This again suggests that growing conditions in the field and/or pathogens present in the growing environment are killing off the cumin.

Table 2.
Effect of seed treatment on plant counts for cumin in the greenhouse – 2005.

Seed treatment Average germination out of 10 seeds
  22 days after planting 12 weeks after planting
L1269 8.2 a 8.6 a
Allegiance FL 8.0 a 7.4 a
Gaucho 480 FL + L1269 6.1 b 6.7 a
Gemini 6.7 ab 7.8 a
Vitaflo 280 7.7 a 8.0 a
Maxim 480 FS 7.6 a 7.6 a
Apron Maxx RTA 6.7 ab 7.5 a
Control (no chemical) 7.7 a 8.0 a
Coefficient of variance 20.0 24.4
* values followed by the same letter in the same column are not significantly different using Duncan’s Multiple Range test, P= 0.05.

Conclusions:
The observation that none of the chemical treatments affected plant counts when compared with untreated seed grown in potting soil suggests that any problems seen in the field are likely due to pathogens present in the soil, not on the seed.  These problems would be exacerbated by any stresses caused by unsuitable field conditions (soil moisture and/or temperature).

2006 trial

The 2005 greenhouse trial indicated that the cause of the most of the problems with stand establishment in cumin and coriander was soil-borne pathogens and/or less than ideal abiotic conditions for seed germination.  The 2006 greenhouse trial was designed to further test this hypothesis

Materials and Methods:
The treated seed lots used in the field trials were again evaluated under greenhouse conditions. Ten cumin seeds of each seed treatment were planted in four replicate 4” pots and the germination counted after approximately two weeks and 10 weeks.  Rather than utilizing just soil-less mix as was done in 2005, we opted to also use field soil from the test sites employed in 2006 (Innovation Place and Potato Research sites). The soils were mixed 1:1 with Sunshine Potting mix #4.  Both sterilized and non-sterilized soils were used.  The soils were sterilized by autoclaving for 30 minutes. Pure Sunshine Mix #4 was also tested. The greenhouse tests were started at the end of June, and the last stand counts were taken in September.

Results
When treated seed was planted into non-autoclaved field soil from both sites there were significant fungicide treatment effects on the number of seedlings established after two to four weeks (Table 3).  In both the Potato Site and Innovation Place soil Apron Maxx and L1269 significantly increased seedling number relative to untreated seed. Other treatments in non-sterilized soil were not significantly different from the control and although Gaucho alone had significantly lower seedling number than Gaucho plus L1269 in the potting mix none of the treatments significantly differed from the control. Apron Maxx and L1269 both contain the active ingredient metalaxyl. This chemical is
particularly effective in controlling Phycomycete pathogens such as Pythium and Phytophthora species, suggesting that one of these pathogens may be responsible for the decline in seedling numbers observed in non-sterilized soil.

Table 3.
Germination and establishment of treated cumin seed in potting mix,
sterilized and non-sterilized field soil in the greenhouse, 2-4 weeks after planting.

Seed Treatment Average no. of plants per pot (10 seeds planted):
- Potato Field soil Innovation Place field soil Sunshine Mix # 4
- Non-sterilized Sterilized Non-sterilized Sterilized
Apron Maxx RTA 6.5 a* 5.0 9.8 a 7.3 5.5 ab
L1269 6.5 a 5.5 9.8 a 8.3 4.8 ab
Gaucho 480 FL 3.3 ab 5.0 1.0 c 8.0 3.0 b
Gaucho 480 FL + L1269 4.5 ab 5.0 6.3 b 6.0 6.5 a
Maxim 480 FS 3.0 b 6.0 4.3 bc 6.8 4.8 ab
Control (no chemical 2.3 b 7.8 3.5 bc 7.8 5.3 ab
Mean 4.4 5.7 5.8 7.4 5.0
Coefficient of variance (%) 48.0 40.3 36.1 31.1 41.9
*values within a column followed by the same letter are not significantly different using Duncan’s Multiple Range test, P = 0.1.

Mean values for the soil types indicated that overall seedling survival was greatest in the Innovation Place soil, and that sterilizing the soil increased seedling survival in both field soils. When the soil was sterilized there was no advantage to using a seed treatment, suggesting that pathogens play a significant role in reducing seedling survival. The extremely poor survival of seedling in the Innovation Place plots, compared with the greenhouse values, implied that factors other than pathogenic attack were playing an important role in reducing the cumin stand in the field – at least early in the growing season.

The numbers of seedlings present in the pots after nine to ten weeks was very similar to the counts after two to four weeks (Table 4). In some cases a few seedlings died and in others late seedlings emerged. In the non-sterile soils, seed treated with Apron Maxx and L1269 still survived significantly better than untreated seed. In the potting soil and sterilized soils there were no significant effects of fungicidal seed treatments, with the exception of a reduction in seedling number for Gaucho plus L1269 compared with the control in sterilized Innovation Place soil.  Treatment with Gaucho alone generally reduced seedling number in all soils, suggesting a phytotoxic effect.

Table 4.
Germination and establishment of treated cumin seed in potting mix,
sterilized and non-sterilized field soil in the greenhouse, 9-10 weeks after
planting.

Seed Treatment Average no. of plants per pot (10 seeds planted):
  Potato Field soil Innovation Place field soil Sunshine Mix # 4
Apron Maxx RTA 7.3 a* 6.8 9.3 a 8.3 ab 6.0
L1269 7.5 a 7.3 7.5 ab 9.0 a 6.3
Gaucho 480 FL 3.3 b 6.8 1.0 d 9.3 a 4.0
Gaucho 480 FL + L1269 4.3 b 6.8 6.5 b 5.8 b 6.5
Maxim 480 FS 3.3 b 8.0 3.8 c 6.5 ab 5.5
Control (no chemical 2.3 b 7.8 3.3 c 8.8 a 6.0
Mean 4.7 7.3 4.7 8.0 5.7
Coefficient of variance (%) 42.7 28.6 26.4 21.1 37.1
* values within a column followed by the same letter are not significantly different using Duncan’s Multiple Range test, P = 0.1.

The mean values again showed that seedling survival after 9-10 weeks was best in sterilized field soil, and was close to the 86% seed germination recorded in sterile petri dishes in the laboratory. Seed planted into Sunshine mix #4 survived better than in the unsterilized field soil but not as well as in the sterilized field soil. This mix does not support seedling survival as well as Sunshine mix #3, which was used in the 2005 tests (average survival of 77% of the seedlings after 3 weeks in the control treatment). Future tests will include autoclaved potting soil mix to determine if this affects seedling survival.

Conclusions:
In 2005 field trials there was a significant advantage to applying seed treatments to cumin seed but in 2006 use of seed treatments did not improve stand establishment under field conditions.  In 2005 seed treated with a mixture of Gaucho and L1269 showed significantly higher seedling survival compared with untreated seed. This effect was seen again in 2006 but the effect was not significant. The variability in the results emphasizes the importance of testing at different sites over a number of years.

Irrespective of seed treatment, the plant stands were very poor in 2006, particularly at the Innovation Place site. This site was not irrigated but soil moisture conditions were adequate throughout June. It is not known if the initial low plant counts were due to non-germination of the seed, or death of seedlings prior to emergence.  These results suggest that applications of seed treatments alone will not solve the establishment problem for the cumin crop.

Seed treatments containing metalaxyl were effective in increasing seedling survival in non-sterilized field soil used in greenhouse experiments. The greenhouse trials were conducted during the summer months when the seed would be germinating and growing at daytime temperatures around 23°C. It is possible that the spring conditions in Saskatchewan are too cold for the seedlings to become established, particularly if there is disease pressure from root rotting pathogens such as Pythium and Phytophthora species.  These pathogens are competent over a wide temperature range.  A later planting of cumin in the field at Scott Research Farm in the summer of 2006 appeared to survive much better than early spring planting  (D. Ulrich, personal communication). Unfortunately cumin is susceptible to frost in the fall so later planting may not allow sufficient time for the crop to mature.

Greenhouse trial 2007

In 2007 a similar greenhouse trial was conducted to repeat the evaluation of the effect of seed treatments on seedling establishment reported in 2006. Ten cumin seeds of each seed treatment were planted in four replicate 4” pots and the seedling emergence counted after two and four weeks. In this trial three field soils were used, along with potting soil alone. The soils were from the University Potato Research site (University), from a potato farm in the Allan area (Allan) and from a spice growing farm in the Eston area (Eston). These soils were all a light, sandy type of soil, and cumin had been grown at the University and Eston sites. The soils were mixed 1:1 with Sunshine potting mix #4 (Potting mix), and both sterilized and non-sterilized soils were used.  A sterilized potting mix treatment was included along with a non-sterilized potting mix treatment.  The trial was set up as a factorial design, unlike the previous trial when there were unequal treatments as a sterilized potting soil treatment was not included.

The seed treatments were applied as in the previous trial and were:

1. Apron Maxx RTA (Syngenta), a.i. 0.73% fludioxonil, 1.10% metalaxyl-M,
applied at 9.6 ml product/kg seed

  1. Trilex (formerly L1269, Bayer CropScience), a.i. 10.8g/L metalaxyl, 13.5g/L trifloxystrobin, applied at a rate of 7.4 ml product/kg seed
  2. Maxim 480FS (Syngenta), a.i. 40.3% fludioxonil, applied at 0.21 ml product/ kg seed
  3. Control, seed treated with 10 ml water.

The Gaucho treatments were not included in this trial as the insecticide appeared to have a phytotoxic effect in 2006, which could confound the effect of the fungicide. The trial was started in mid November and terminated at the end of December.

Results and Discussion

In the first seedling count after 2 weeks, there was no effect of seed treatment but there was a significant effect of soil type, and an interaction between soil type and soil sterilization (Table 5). Significantly more seedlings emerged in all the field soils after sterilization, suggesting that pathogens may be responsible for reducing emergence in these soils. Significantly more seedlings emerged in the Allan soil than in the other two field soils. This implies that sterilization of the Allan soil was particularly effective in removing some factor that was causing reduced emergence in this soil. This factor may be a pathogen found specifically in the Allan soil but not at the other two sites. However, the seed treatments had no effect after two weeks, suggesting this factor is not controlled by fungicides. There was no effect of sterilizing the potting soil on seedling emergence, as the non-sterilized potting mix had a very high rate of seed emergence.

Table 5.
The effect of soil type and soil sterilization on the establishment of cumin seeds two weeks after planting in the greenhouse.

Soil type No. of seedlings emerged (out of 10 planted)
- Non- Sterilized soil Sterilized soil
University  0.95 c* 4.75 b
Allan 1.30 c 6.95 a
Eston 2.20 c 4.70 b
Potting  mix 8.00 a 7.05 a
* values followed by the same letter within the table were not significantly different using the Student’s t test at p<0.05.

After four weeks, seedling emergence had increased overall, from an average of 4.5 to 7.0 seedlings per pot.  Emergence was significantly affected by the soil type and by sterilizing, and there was again an interaction between soil type and sterilization (Table 6).  Again there was no difference between sterilized and non-sterilized potting soil.   After four weeks there was also no difference in seedling establishment between the sterilized field soils and the potting soil.

It also appeared that after four weeks the non-sterilized Eston soil was less detrimental to seed establishment than the other two field soils, although the stand was still significantly lower than in the potting soil mix.

Table 6.
The effect of soil type and soil sterilization on establishment of cumin seeds four weeks after planting in the greenhouse.

Soil type No. of seedlings emerged (out of 10 planted)
  Non- Sterilized soil Sterilized soil
University   3.85 c* 8.60 a
Allan 3.65 c 8.25 a
Eston 5.60 b 8.90 a
Potting  mix 8.40 a 8.65 a
* values followed by the same letter within the table were not significantly different using the Student’s t test at p<0.05.

After four weeks there was a significant interaction between the seed treatment in the different soils (Table 7) and between the sterilization treatments of the different soils (Table 8), at the 10% probability level.

Table 7.
The interaction of soil type and seed treatment on establishment of treated cumin seeds four weeks after planting in the greenhouse.

Soil type No. of seedlings emerged (out of 10) in the different  seed treatments
  Apron Maxx Trilex Maxim Control
University 7.40 bc* 6.50 cde 5.8 def 5.20 f
Allan 6.70 cde 6.50 cde 5.00 f 5.60 ef
Eston 6.90 cd 7.20 c 7.40 bc 7.50 bc
Potting mix 8.40 ab 8.40 ab 8.40 ab 8.90 a
Mean 7.3 7.1 6.7 6.8
* values followed by the same letter within the table were not significantly different using the Student’s t test at p<0.10.

In the University soil, treatment with Apron Maxx and Trilex significantly increased seedling number compared with the untreated seed.  The Apron Maxx treatment was the only treatment where the seedling numbers in the University soil were not significantly lower than in the potting mix.  When seed was treated with Maxim, seedling number in the Eston soil was not significantly different than in the potting mix, but seedling numbers were not significantly higher than for untreated seed.  This indicates that the seed treatments have differing efficacies, depending on the soil type. This is consistent with the interaction observed between seed treatment and soil sterilization (Table 9).

Seedling numbers were consistently higher in sterilized versus non-sterilized soils for all three field soils but there was no effect of sterilizing the potting mix ((Table 8). There were significantly more seedlings produced in the non-sterilized Eston soil compared with the other two non-sterilized field soils, suggesting the Eston soil has fewer or less virulent pathogens present that are controlled by sterilization, as there were no differences in seedling establishment when all the soils were sterilized.

Table 8
The interaction of soil type and soil sterilization on establishment of cumin seeds four weeks after planting in the greenhouse.

Soil type No. of seedlings emerged (out of 10 planted)
  Non- Sterilized soil Sterilized soil
University   3.85 c* 8.60 a
Allan 3.65 c 8.25 a
Eston 5.60 b 8.90 a
Potting  mix 8.40 a 8.65 a
* values followed by the same letter within the table were not significantly different using the Student’s t test at p<0.05.

In the non-sterilized soil, treatment with Apron Maxx increased seedling establishment relative to untreated seed and the other two fungicide treatments (Table 9). Apron Maxx has shown efficacy in increasing seedling establishment in previous trials, and would therefore be the seed treatment most likely to be effective in the field. However the efficacy would depend on which pathogens were present in the field soil.  Apron Maxx contains two fungicides that are effective against a broad spectrum of pathogens such as Pythium and Phytophthora species and Fusarium and Rhizoctonia species. In the sterilized soil, treatment with Trilex increased establishment, but the other two fungicides did not affect establishment in comparison with the untreated seed.  Trilex appeared to have some growth promoting activity unrelated to pathogen control as positive responses to this treatment were only seen in the sterilized soil. Trilex was not effective in non-sterilized soil, suggesting that this treatment may not be effective against the pathogens causing seedling blight in cumin in the field.

Table 9.
The interaction of seed treatment and soil sterilization on establishment of treated cumin seeds four weeks after planting in the greenhouse.

Seed treatment No. of seedlings emerged (out of 10 planted)
- Non- Sterilized soil Sterilized soil
Apron Maxx 6.45 c* 8.25 b
Trilex 5.05 d 9.25 a
Maxim 4.65 d 8.65 ab
Control 5.35 d 8.25 b
* values followed by the same letter within the table were not significantly different using the Student’s t test at p<0.05.

Conclusions:
Seedling emergence and establishment were affected by the type of soil they were grown in, and when the soil was sterilized seedling emergence generally increased. Seed treatments did not affect initial seedling emergence but after four weeks Apron Maxx treatment had an effect in one of the soils. This indicated that different soils contained different pathogens that responded differently to the seed treatments.
Overall conclusions for efficacy of seed treatments on cumin stand establishment in the greenhouse:
Every soil type is likely to have a specific microflora of pathogens that respond differently to available seed treatments. Thus no single treatment will be effective in all soils. The problem of seedling emergence and stand establishment of cumin is complex and requires more research into understanding the factors responsible for the poor stands observed under Saskatchewan conditions.

4. Efficacy and Crop Tolerance Trials for Azoxystrobin for Control of Blossom Blight in Coriander (Coriandrum sativum)

Blossom blight appears to be a common problem in coriander. The causal organism for blossom blight of coriander appears to vary from region to region. In Saskatchewan, Duczek (2002) isolated a range of potentially pathogenic fungi from affected plants, but identifiedAureobasidium sp. as the primary pathogen in a series of controlled environment trials.

Food Blossom Blight
As the name implies, blossom blight typically attacks the flowers, although in severe cases leaves and stems adjacent to heavily infested flowers may also be damaged. Infected flowers fail to set seed resulting in substantial yield loss. The primary inoculum source for blossom blight is likely wind or water borne spores from infected crop residues from the previous season. Establishment and spread of blossom blight is promoted by rain and/or irrigation but in dense canopies dew may be sufficient to allow establishment and localized spread. Commercially available germplasm appears susceptible to blossom blight. Chemical control therefore represents the next line of defence. In preliminary trials conducted by Duczek (2002) and Waterer (2003) a range of foliar applied fungicides appeared to provide at least some protection against blossom blight of coriander and other spice crops. Azoxystrobin (PCP # 256153) applied at first flowering appeared to be one of the more promising products.

This trial is designed to provide crop tolerance and efficacy data in support of minor use registration for azoxystrobin for control of blossom blight in coriander. In addition to the treatments stipulated by the minor use testing program a range of other treatment combinations were evaluated.

Materials and Methods
The trial was conducted at the University of Saskatchewan, Horticulture Field Research Center in Saskatoon, SK. This site features a Sutherland series clay soil, (pH 8.0, EC< 1.0 dS, CEC 41 meq/100g, 4.4% organic matter, 65% clay, 25% silt, 10% sand). This site has been used in for varietal development and disease screening on a range of spice crops including coriander for the past three growing seasons. In both 2002 and 2003 blossom blight had been observed in coriander planted at this site.

The plot area was disked and harrowed prior to planting. The registered herbicide Treflan (trifluralin @ 1L ai/a) was applied prior to the field preparation step. No fertilizers were applied as soil tests indicated the plot area had adequate levels of all nutrients (110 kg N/ha, 175 kg P2O5/ha, >1000 kg K2O/ha). Coriander seed (cv. CDC Major) to plant this trial was obtained from Schweitzer’s Seed (Gary Schweitzer) of Eston SK. The supplier had not observed any problems with disease in the year the seed was produced. The plot was seeded on May 5 using a John Deere disk drill. The seed was planted to a depth of 4 cm with 20 cm between drill rows. The seeding rate was relatively high (25 kg/ha), as a complete/thick stand tends to promote development and spread of disease in caraway.

In the last week of June we implemented a supplemental irrigation program in an effort to promote both crop development and the onset of conditions conducive to disease. Each week the plot was irrigated for a minimum of 1 hour (1 cm water applied) utilizing a wheel-move type irrigation system. On June 30, the registered herbicide Afolan F (linuron @ 0.5 L ai/a) was applied to control emerged broadleaf weeds. The herbicide program provided a good level of weed control – and weed competition was not an issue in this trial.

In early July, individual treatment plots were created by tilling out 0.6 m strips in the main plot. Each individual treatment plot measured 10 m * 1.5 m. A randomized complete block design was utilized with four replicates.

Fungicide treatments
The treatments tested in this project were:
  • 1. Quadris (Syngenta) at 55, 109 or 280 g ai (azoxystrobin)/ha, single application, applied at early bloom.. This combination of rates and time of application reflects the protocol specified in the minor use testing program.
  • 2. Bravo (Syngenta) applied at 1250 g ai (chlorothalonil)/ha, single application, at early bloom – represents the treated check. This application rate was based on recommendations from the manufacturer for control of similar foliar disease problems in similar stature crops.
  • 3. Quadris applied at 250g ai/ha at two week intervals beginning at bloom
  • 4. Dithane (mancozeb, Rohm & Haas) applied at 1800g ai /ha beginning at bloom and repeated on Aug 17 + Bravo applied at 1250g ai/ha on Aug 1
  • 5. Headline (pyraclostrobin BASF Canada) applied at 100g ai /ha beginning at bloom and again two weeks later + Lance (boscalid, BASF Canada) applied at 300g ai/ha (Aug 17)
  • 6. Lance applied at 300 g a.i./ha at two week intervals beginning at bloom and running through Aug 17.
  • 7. Control – no chemical applied – represents the untreated check.

The spray program was initiated on July 26th at which time the crop was just beginning to come into bloom. The sprays were applied using a CO2 powered backback sprayer (276 kPa) utilizing cone-type nozzles. The fungicides were applied in the equivalent of 200 l water/ha.

Disease Evaluation
The first disease rating was conducted on August 6th , 11 days after the fungicide treatments were applied and 7 days after symptoms of disease were first observed. Disease incidence was assessed by examining 10 randomly chosen locations per plot for presence/absence (+/-) of blossom blight. The number of infected areas per plot was evaluated again on Aug 19 and 26. The final disease assessment was conducted on September 9. At that time, the % of each plot affected by blossom blight was evaluated.
Harvest
Each plot was individually harvested by direct combining on October 7. Recovery of the coriander fruit (seeds + associated ovary elements) was excellent. The fruit were dried at 30oC for 48 hours and then were run through a dockage tester (Carter and Day – Model # XT3) to remove debris and to segregate out the fruit that had failed to set seed. Yield of both empty and full fruit balls was determined for each plot. The 1000 seed (fruit) weight was determined for each treatment replicate.

Results
-
Weather Conditions and Crop Observation
Below normal temperatures delayed emergence of the coriander crop until the 2nd week of June (5 weeks after seeding). The cool conditions also resulted in a staggered emergence pattern – with plants continuing to emerge through early July. This resulted in non-uniformity of crop staging for the duration of the growing season. Replicates 2 and 4 were clearly more developmentally advanced than the other replicates. This contributed to the significant block effects seen in all statistical analyses.

By late June weather conditions were more favorable and the crop showed decent vegetative growth. However, the remainder of the 2004 cropping season was abnormally cool and cloudy and crop development was slow. The crop had begun to bolt by the 2nd week of July, with the first flowers opening a week later. By the 2nd week of August, the coriander plot was in full bloom and because of the cool, moist conditions it continued to bloom until early September. On August 19th the plots were hit by a light frost (-2C) but no crop damage was observed. A killing frost occurred in the last week of September.

No insect pests were observed. Sclerotinia and aster yellow diseases were observed at low levels late in the season – but these diseases were of limited severity relative to the blossom blight. There was some lodging, but this did not interfere with crop recovery at harvest.

Weather conditions at the time of spraying (July 26th – 7 am ) were calm and 15C. At 12:30 pm that day a brief rain shower occurred. There were also light showers on the 27th and 28th … but the total accumulated moisture was minimal (0.10 cm) and should have not compromised product efficacy.

- Disease Development
The first signs of blossom blight were observed on July 29, with plants in one area of the plot showing the characteristics symptoms of browning of the developing flowers and associated leaves. This affected area slowly increased in size, eventually covering ca. 10 m2. Additional diseased areas appeared in the plots daily and these disease zones also increased in size.

- August 6 – the disease was just getting established in the plots at this time. Disease distribution was non-uniform both within blocks and across treatments. There were no significant differences between disease incidence in the controls versus any of the spray treatments at this time.

- August 19 – disease levels had increased relative to the previous sampling date, but the disease distribution continued to be non-uniform across both treatments and blocks. There were no significant differences between disease incidence in the controls versus any of the spray treatments at this time.

- August 26 – averaged over the entire trial area, disease was found in over 75% of all sites examined. There were no significant differences between disease incidence in the controls versus any of the spray treatments at this time.

- September 9 – the plots were rated as to the % of the total plot showing browning and blighting of flowers – no differences were observed between any of the fungicide treatments and the control. On average 36% of the plot area was affected by disease at this time.

 

Table 2.1.
Means and analyses of variance for fungicide effects on coriander disease ratings at various points in the growing season.

Treatment Average number of infected spots
(out of 10)
% plot area infected
  August 6 August 19 August 26 September 9
Quadris - - - -
-55 g ai/ha 0.5 2 7 40
-109 g ai/ha 0.25 1 8.2 35
-280 g ai/ha 0.25 1.25 8 32
Bravo (1 applic) 0 0.25 7.8 40
Quadris (3 applic) 0.25 1.75 6.8 35
Dithane/ Bravo/ Dithane 0 2 7 29
Headline/ Headline/ Lance 23 0 5.5 29
Lance (3 applic) 0 1 8.8 22
Control 0 1.8 0.2 32
P values for treatment 0.44 0.32 0.66 .09

Yields
There were no significant treatment effects for any of the yield parameters measured. Block to block variability in yields was high … in part due to uneven crop development and in part due to the uneven distribution of disease within the plots. Many of the fruit balls failed to form seed, particularly in areas hardest hit by the blossom blight. Dockage rates exceeded 40% for all treatments.

There were no indications (visual or yields) of any of the fungicide treatments exerting phytotoxic effects on the crop.

Conclusion
None of the fungicide treatments tested resulted in disease levels that differed significantly from untreated check areas. Coriander is an indeterminate type plant with an extended bloom period. Based on the results form this single year of testing, it does not appear that a single application of a protectant type fungicide such as Quadris has much potential to protect the crop from blossom blight. Although multiple applications would likely be required – none of the treatment combinations tested in 2004 provided a significant degree of control of ths disease problem.

/you have reached the end of Objective 2 - Click here to return to the top of the page

-

OBJECTIVE 3
Improved agronomic practices for milk thistle.
Milk Thistle Swws
Milk Thistle Seed
Flowering Milk Thistle
Flowering Milk Thistle

MILK THISTLE BACKGROUND
Milk thistle (Silybum marianum (L.) is a flowering plant belonging to the daisy family (Asteraceae).  Milk thistle is known by several other names, including Blessed Milk thistle, Spotted thistle, St. Mary’s thistle, Marian thistle, Holy thistle and Variegated thistle. Milk thistle is native to the Mediterranean regions of Europe, North Africa and the Middle East but has been introduced throughout the world. The plant is valued for its medicinal properties, but it is sometimes grown as an ornamental because of its unusual leaves.  Milk thistle spreads quickly and is considered a weed in some parts of the world.In the U.S.A., it grows wild in most southern states and the north-eastern and mid-west states. It has been declared a noxious weed in Washington, Oregon and Texas, but is not considered as a noxious weed in any Canadian province.

Plant Description:
Milk thistle is a vigorous tall upright plant that prefers dry sunny conditions.  The spiny stems branch at the top, and reach a height of approximately 1.2 to 2 m. The spiny leaves are wide, with white blotches or veins. Milk thistle gets its name from the thistle like appearance of the leaves and the milky white sap that exudes from any cut surface.  Flowers are generally red-purple.  A solitary flower develops at the end of each stem.  Milk thistle has an indeterminate growth and flowering habit, resulting in uneven development and maturity of flower heads.  The mature fruit (achene) is relatively small with an attached white silky pappus.  In the literature, the achenes, i.e. the ‘fruit’ are mostly (wrongly) referred to as ‘seed’. The immature fruit is soft and cream, tan, or brown in colour and the mature fruit is hard-skinned, shiny brown or brown with tan spots.  Due to the indeterminent flowering habit at any point in time the milk thistle plant has some flowers that are still opening while others have progressed to the point where the seed is shattering.  To address this issue of uneven seed maturity the mature heads are usually gathered by selective hand harvest.  The large thorns on the stems, leaves and seed heads of milk thistle make hand harvest of the seed heads an exceedingly unpleasant task.   While once-over machine harvest is clearly preferable from the perspective of efficiency, seed yield and quality may be compromised.

In the relatively cool and short Saskatchewan growing conditions, milk thistle is grown as an annual while in warmer longer growing environments milk thistle can be grown as a biennial.

Uses of Milk Thistle:
Every part of milk thistle including the stems, leaves, flower buds are edible. Milk thistle has been used for over 2,000 years as a herbal remedy for a variety of ailments, particularly for liver and gall bladder problems.  Studies suggest that substances in milk thistle (especially a flavonoid called silymarin) protect the liver from toxins, including certain drugs such as acetaminophen (Tylenol), which can cause liver damage in high doses. Silymarin has antioxidant and anti-inflammatory properties, and it may help the liver repair itself by growing new cells.

The leaves are used as a salad green or cooked.  Leaves can be trimmed of prickles and boiled or added raw to salads.  There are as yet no known medicinal properties for milk thistle leaves.  It is claimed that in recent times a number of herbal medicine manufacturers have introduced products containing milk thistle leaves but these products had no apparent therapeutic value.

Constituents:
The principal extract of milk thistle fruit, silymarin (4% to 6% in ripe fruit), is composed of several polyphenolic flavonolignans. The major component (60%) is silybin (also known as silibinin or silybinin) is also the most biologically active component.  Other components include silichristin (also known as silychristin, silycristine or silicristin), a metabolic stimulant, and silydianin. Silymarin is found in highest concentrations in the fruit of the plant. Other constituents are flavonoids, a fixed oil (16% to 18%), betaine, trimethylglycine (TMG) and amines.


Silybin

Silichristin

MILK THISTLE GERMPLASM TRIALS IN 2005

Methodology:
Lines tested in 2005 were; a) adapted lines tested in 2004 and b) any additional lines received.

The field test was conducted at the Department of Plant Sciences, Horticulture Field Research Station in Saskatoon.  This site was found to be well suited to production of milk thistle – based on the 2004 trial results.

The trial site was prepared by disking, then rotovating two weeks prior to seeding. As a function of heavy fertilizer applications in previous crops, fertility levels in the test area were > 100 kg N/ha and > 120 kg P2O5/ha.  These fertility levels would easily meet the needs of most herbaceous medicinals and may actually be somewhat excessive – potentially leading to excessive vegetative growth.   Seeding occurred earlier in 2005 than in 2003 or 2004 (May 6) in an attempt to maximize the length of the growing season.  The plots were seeded using a push-type precision small plot seeder.  Rows were spaced 0.75 m apart with 0.75 m between plants within the row.  There were two rows in each plot giving a total plot area of 12m2.  There were two replicates arranged in a randomized complete block design.

Emergence percentages for most lines were around 30%.  As the crop was seeded relatively heavily, most lines had to be hand thinned to prevent overcrowding.  Emergence rates of Reddis and Cruz E were poor despite obviously favorable germination conditions.  This suggests problems with seed quality -   particularly as these lines also had emergence problems in previous years.

As milk thistle is slow to establish, some weed control was needed.  However, the wide row and plant spacings made tillage simple. No problems with weed competition occurred in 2005 – although there was obviously competition between milk thistle plants within the row.  No herbicides were applied in this trial in order to avoid confusing agronomic performance with issues of herbicide sensitivity.

Although growers have reported problems with Aster Yellows in Milk thistle, no problems with disease were observed in this trial.  A heavy infestation of Painted Lady Butterfly occurred throughout Saskatchewan in 2005.  Milk thistle is one of the preferred food sources for this occasional pest.  The larvae began to emerge in late June at which time the plants were about 30 cm tall and growing vigorously.  Although the larvae caused significant defoliation – their lifespan is limited and it is doubtful that the feeding damage would have any longterm impact on a plant as vigorous as milk thistle.  Nonetheless – we sprayed the plots with insecticide (Decis) is late June.  A single application provided excellent pest control.  Organic products like Bt. would likely provided adequate control – if problems with Painted Ladies become common.

Note – Milk thistle growers in other areas of the province lost their crop to Painted Lady Butterflies in 2005.     

The plot area was not irrigated in 2005.  This reflects the fact that; a) 2005 was fairly cool and wet and b) results from previous years suggested limited need or benefit of irrigation.  Although there were indications of depletion of soil water by late August, the plots were not irrigated in an effort to promote maturity.

By late-August some of the lines were approaching maturity.  On Sept 5 the plots were sprayed with the chemical desiccant Reglone (1 l/a) in 100 l/a of water.  Although most lines were still growing vigorously at the time of desiccation – this single Reglone application provided excellent top kill.  All the tops were dead within 3 days of treatment.  The plots were harvested on Sept 8, which is one month earlier than the harvest in 2004.  The entire trial was harvested using a Wintersteiger Nursery Master Elite small plot combine.  The header was raised to minimize the volumes of vegetative material taken in along with the seed heads.  The combine was adjusted to minimize seed losses by reducing the fan speed and opening the sieves.  The harvesting operation was both quick and efficient.  The majority of seed losses reflected shattering prior to the harvest, rather than seed missed by the combine.  Few of the lower canopy heads missed by the header contained any useful seed.  The harvested material was air dried at 40 C for 5d, threshed and then cleaned using a dockage sorter.   Seed yields and 1000 seed weights were determined at this point.  For the quality analysis a seed sample was further cleaned by hand to eliminate all impurities.

The stand, maturity index and yield data for 2005 again show that some milk thistle lines were far better suited to SK growing conditions than others (Table 1).  Although 2005 was intermediate in temperature between 2003 (hotter) and 2004 (cooler), the 2005 crop was higher yielding with more advanced and uniform maturity than either the 2003 or 2004 crops.  This reflects both the growing season and improvements in production practices.   One line (Genesis seeds) was overmature by the time of the harvest – and losses due to shattering may have been higher than normal in this line (* in table).  As noted in previous trials, some lines were just beginning to flower at the time of the final harvest.

When seed yields are considered;
a) Milk thistle appears to be capable of producing acceptable yields even with limited plant populations.  Once established, the plants grow rapidly and make efficient use of all available growing space.  Consequently, high seeding rates and/or high rates of seed germination  are not required for good yields.
b) several of the lines had seed yields that exceeded 1500 kg/ha.  This is promising as the yield levels needed for milk thistle to represent an economically viable crop in Saskatchewan are ca. 500 kg/ha.  It is important to note that yields in small plot field trials are a questionable indicator of the yields that can be expected under commercial production conditions.  However, the fact that overall yields increased significantly over the 3 years of these trials suggests that the utilization of better management practices and greater familiarity with the needs of the crop, could result in even higher yields.
c) there was again little correlation between seed size and yield

The quality data for the various milk thistle lines tested in 2005 are presented in Table 2, along with the quality data for selected treatments from the agronomy trials.

Some key quality observations from the 2005 germplasm evaluation trials are;
a) The average silymarin content in 2005 was 1.5% compared with 1.4% in 2004 and 1.6% in 2003.  It is interesting to note that the silymarin content appears to reflect the trend in growing season temperatures … 2003 > 2005 > 2004.   The silymarin content of milk thistle seed is known to increase as the seed matures.  These data support that concept and points out the need to utilize cultivars and production practices that maximize seed maturity within the available cropping season.
b) relative seed quality of the cultivars was not overly consistent from year to year.  For example, although there were many cultivars that had more than 2% silymarin in a given test year, none of the cultivars exceeded 2% silymarin across all three test years.  This lack of consistency will represent a major challenge if the marketplace focuses on a high silymarin content as an indicator of quality.  At present, the market seems to either not consider silymarin content or is content if the crop exceeds some minimum level ie; 1.6%
d) total silymarin yield (seed yield * silymarin content) is another potential measurement of a cultivar’s value … assuming that silymarin content is a factor in sale price.   In the 2004 trial, Austra Hort, Bolier and Kalyx-1 were outstanding for total silymarin yield.  In the 2005 trial,  Austra Hort and Kalyx-1 were again amongst the top ranking lines for total silymarin yield.  Bolier had only mediocre yields and seed quality in 2005.  The highest yielding line in 2005 (Royston Petrie 131) was not tested in 2004 … but had performed well in 2003.

Table 1.
Growth and yield characteristics for Milk Thistle lines in 2005.

Line Country of 
Origin
Emergence  Yield
 (kg/ha)
1000 seed wt
 (g)
Austra Hort Holland 33 1805 25.4
Bolier Holland 40 1369 25.7
Cruz Eastern Europe 25 1253 27.2
Genesis Israel 42 673* 32
Horizon Herbs ? 33 918 22.9
Johnny’s ? 30 1050 23.1
Kalyx 1 Holland 57 1708 26.6
Reddis India 9 1052 27.9
Richters ? 50 1204 22.7
Royston
Lot 1236 LJ2
Germany 46 1541 23.2
Royston
Lot 131XJ2
Germany 41 1887 24.9
Royston
Lot 757PJ2
Germany 79 1371 23
TMP 14068 ? 65 1670 24.8
AVG - 42 1346 25.3

 

Table 2.
Quality characteristics for Milk Thistle lines evaluated in 2005

Line Silibinin A
(% w/w)
Silibinin B
(% w/w)
Other Flavonolignans Silymarin 
(% w/w)
Silymarin Yield (kg/ha)
Austra Hort 0.3 0.41 0.95 1.67 30.1
Bolier 0.21 0.28 0.74 1.23 16.8
Cruz 0.44 0.64 1.02 2.09 26.1
Genesis 0.08 0.11 1.17 1.37 9.2*
Horizon 0.07 0.1 1.21 1.38 12.7
Johhny’s 0.07 0.1 1.14 1.31 13.8
Kalyx 1 0.25 0.36 0.84 1.45 24.8
Reddis 0.38 0.54 0.98 1.9 20
Richters 0.27 0.38 0.91 1.56 18.8
Royston Lot 757PJ2 0.4 0.6 0.74 1.74 23.9
Royston Lot 1236 LJ-2 0.14 0.19 0.97 1.3 20
Royston Lot 131 X32 0.3 0.45 0.95 1.7 32.1
TMP 14068 0.08 0.11 0.91 1.11 18.5
Mean 0.23 0.32 0.96 1.52 20.5

NEW MILK THISTLE LINES RELEASED BY THE UNIVERSITY OF SASKATCHEWAN

The objective of this program was to develop lines of milk thistle (Silybum marianum) that were suited to mechanical harvesting. Short stature plants with uniform seed maturity and good seed retention within the seed head are all desirable characteristics in a mechanically harvested crop.   Early maturity, high seed yields with acceptable seed quality were also selection criteria.

Line Strengths Limitations
U of S-1 High yields.  Short stature and early, uniform maturity.  Good total silymarin content and silybinin B levels are above average.  Relatively light seed coat color.
U of S-2   Exceptionally high levels of silybinin B and overall silymarin content.  Matures early.  Represents the only U of S line being released with seed coat darker than industry standard (Richters). Low rates of seed emergence may limit yields.
U of S-3 Above average yields, medium stature, uniform early maturity.  Slightly above average silybinin B levels. Light seed coat color.
U of S-4  Good yields. Above average silybinin B.  Fairly dark seed coat.  Larger, later maturing plant, but still earlier and shorter than the  industry standard (Richter’s).

Table 1.
Characteristics of new U of S Milk Thistle Lines.  Richter
s line of milk thistle is included as a comparative standard. 

Line Stand
(%)
Height
(cm)
Maturity
(%)
Yield
(kg/ha)
Silymarin
(%)
Silybinin B
U of S-1 58 140 >80% 1550 1.85 0.47
U of S-2 25 160 70 1250 2.10 0.61
U of S-3 67 160 75 1300 1.60 0.40
U of S-4 76 180 50 1450 1.75 0.44
Richters 68 200 50 1040 1.70 0.48

All data are averaged over 2 or more crop years.
Maturity = % of heads reaching maturity by the 2nd week of Sept.

All lines have been licensed to the Herb, Spice and Specialty Ag Assoc
Contact : Connie Kehler – Manager of the SHSA for information on seed suppliers 

 

MILK THISTLE AGRONOMY TRIALS (2003-2005)
Jazeem Wahab (CSIDC, Outlook, SK)
Doug Waterer (University of Saskatchewan, Saskatoon, SK)

Background
Milk thistle is native to the Mediterranean, but is now widespread throughout the world. This stout thistle usually grows in dry, sunny areas. The stem branches at the top and reaches a height of about 1.5 to 4 m. The leaves are spiny and wide, with white blotches or veins. The flowers are red-purple. The small, hard-skinned seed is brown, spotted, and shiny.  Milk thistle is an easy to grow annual plant.  It has an indeterminate growth and flowering habit, resulting in uneven development and maturity of flower heads.

A group of flavenoids called “silymarins” represent the medically valuable compounds sought after in milk thistle.  Silymarins have been demonstrated to both protect and alleviate problems with liver and kidney function.  Silymarins are present throughout the milk thistle plant, but occur at the highest concentration in mature seed.  By comparison, immature seed contains relatively little silymarin.  Consequently, the timing of seed harvest is critical.  However optimum harvest timing is complicated by the fact that;
a)  milk thistle is a late maturing plant with an indeterminate flowering habit.
b) milk thistle is prone to shattering once the seeds are mature

At present, milk thistle is harvested by hand-cutting each seed head as it matures.  This process is slow and costly, particularly as the plants sport very large and sharp thorns.  Saskatchewan growers are struggling to effectively compete with other potential suppliers of milk thistle who have access to more available and affordable sources of hand labor.

Once-over mechanical harvest of milk thistle is possible but yields can be less than 50% of those achieved by hand harvesting (Wahab 2002).  The quality of the resulting crop is also poor as the once over harvest combines mature and immature seeds.  The practical and economic viability of mechanical harvesting milk thistle may be improved if …;

a) lines of milk thistle with more uniform flowering habit are identified.   This objective was addressed in the germplasm evaluation component of this report.

b) agronomic practices are identified that enhance uniformity of crop development and flower maturity

c) pre-harvest desiccation is used to accelerate and synchronize maturity of the flower heads.

The objectives of this project include:

i.    Seeding rate and row spacing effects on yield and quality.
ii.   Seeding date effects on growth characteristics and productivity.
iii.  Nitrogen and phosphorus effects on productivity.
iv.  Effectiveness of vinegar as a desiccant for machine harvest compared to Reglone.

Study Description
Agronomic studies were conducted 2003 through 2005 at the Canada-Saskatchewan Irrigation Diversification Center in Outlook Saskatchewan.  All studies were conducted under dryland conditions, as previous experience at CSIDC showed that irrigation can prolong vegetative growth and delay maturity of milk thistle, resulting in considerable yield losses.   The soil at the test site is a clay loam.  The soil properties at the test locations are summarized in Table 1.  Land was prepared in the traditional manner to form a firm seed bed.  Milk thistle seed from Richter’s Herb Co. was used in the studies.  This line was selected as it has produced reasonably good yields in previous trials and has a good quality profile.  Each test plot was 3.7 m x 1.2 m.  Seeding was done with a double-disc press drill and seeds were placed approximately 1-2 cm deep.  Rows were spaced 60 cm apart with seeding rate of 100 seeds/m2 for all studies except for the seeding rate and row spacing study.

Milk thistle initially grows relatively slow but once established it grows vigorously and competed very well against late season weed pressure.   One early weeding was sufficient.   The crop was desiccated when the flower heads matured, i.e. formation of pappus on the seed.  During maturity, the purple petals dried and white pappus appeared; the seed turned from light brown and soft texture to dark brown and hard.

Two desiccants, Reglone (2.7 l/ha in 1000 l water/ha) and vinegar (14% acetic acid at 1000 l/ha) were sprayed for desiccation during the appropriate stages based on the trial, i.e  when flower heads were 30% and 60% mature.  Seeds were harvested using a Wintersteiger plot combine.  

Table 1: Spring soil analyses (0-30 cm depth) at Milk Thistle test sites: 2003-2005

  2003 2004 2005
Soil texture Clay loam Clay loam Clay loam
Soil pH 8.2 8.3 8.3
E.C: 1S:2W (mS/cm) 0.4 1.1 0.3
E.C. Sat. Extract (mS/cm) 0.9 2.4 0.7
Salinity: Non-saline Slightly saline Non-saline
NO3-N: 28 >144 26
P (kg/ha) > 54 70 >108
K  (kg/ha) 311 398 511
SO4-S  (kg/ha) >43 >86 >86

 

2003 Studies

I.  Effects of nitrogen, phosphorus and harvest stage on yield and quality of milk thistle:

Three levels of pre-plant nitrogen (0, 50, 100 kg N/ha.) and three levels of phosphorus (0, 60, 120 kg P2O5/ha) were examined in this study.  Field plots were laid out as a 3 (nitrogen) x 3 (phosphorus) factorial in a Randomized Complete Block Design with four replications.  Two similar trials were conducted for desiccation at two different stages (30% and 60% mature flower heads) using Reglone.  Seeding was done on May 23, 2003.  The first desiccation (‘Early’) was carried out when approximately 50%-60 % of the flower heads were mature (September 2, 2003) and the second desiccation (‘Late’) six days later (September 8, 2003).  The ‘Early” desiccated crop was harvested on September 5, while the ‘Late’ desiccated crop was harvested September 12.

II.  Seeding rate and row spacing effects on yield for milk thistle desiccated with vinegar. 

Six seeding rates (25, 50, 75,100, 125, 150 seeds / m2) and two row spacing (20, 60 cm) were evaluated.  Field plots were laid out as a 6 x 2 factorial in a Randomized Complete Block Design with four replications.  Seeding was done on May 23,  2003.  The crop was desiccated with Vinegar on September 12, 2003 (when 50%-60% of heads were mature) and harvested on September 19, 2003

2004 Studies:

I. Effects of nitrogen, phosphorus on yield and quality of milk thistle when desiccated with vinegar and Reglone at two maturity stages:

Three levels of pre-plant nitrogen (0, 50, 100 kg N/ha.) and three levels of phosphorus (0, 60, 120 kg P2O5/ha) were examined in this study.  Field plots were laid out as a 3 (nitrogen rate) x 3 (phosphorus rate) factorial in a Randomized Complete Block Design with four replications.  Four similar trials were conducted for the combination of the two desiccants (vinegar and Reglone) and two stages of crop maturity (30% and 60% maturity).  Seeding was done on May 14, 2004.  The ‘Early’ desiccation (30% maturity) was carried out on September 7, 2004 for Reglone, and on September 8, 2004 for vinegar.  The ‘Late’ desiccation (60% maturity) was carried out on September 15, 2004 for both vinegar and Reglone.  Seed harvest was taken on September 27, 2004.

II. Seeding rate and row spacing effects on yield for milk thistle desiccated with vinegar. 

Six seeding rates (25, 50, 75,100, 125, 150 seeds / m2) and two row spacings (20, 60 cm) were evaluated.  Field plots were laid out as a 6 x 2 factorial in a Randomized Complete Block Design with four replications.  Seeding was done on May 14, 2004.  The crop was desiccated with Vinegar on September 15, 2004 (when 50%-60% of heads were mature) and harvested on September 24, 2004.

2005 Studies:
Studies conducted during the summer of 2005 were similar to those conducted in 2004.  Test plots were seeded on May 13, 2005.  The ‘Early’ desiccation (30% maturity) was performed on August 27, 2005 and the ‘Late’ desiccation (60% maturity) was performed on September 1, 2005.  The seed crop was harvested on September 19 and 20, 2005.
Fall Seeding Study:
Studies were also conducted in 2003 and 2004 to examine the feasibility of seeding milk thistle during the fall.  The objective of fall seeding was to advancing flowering and consequently harvest period.  Identical seeding rate and row spacing combinations similar to spring seeding were used for the fall seeding studies.
Results and Discussion
Cropping conditions at the site were generally suited to milk thistle production.  Crop health was good is all years.  The 2003 growing season was relatively warmer and dryer than 2004 or 2005 (Figures 2 and 3).  The 2003 growing season received 133 mm of rain compared to 304 mm in 2004 and 391 mm in 2005.

On average, the 2004 season produced higher seed yields relative to 2003 or 2005.  The 2005 crop had the potential for excellent yields but a heavy rain and wind storm on September 11  (69.2 mm) caused substantial seed loss from the mature heads.  Yields in 2005 were consequently poor.


Figure 2.   In-season rainfall during the 2003, 2004, and 2005 growing seasons.

Minimum Temperature

Average maximum temperature (a) and minimum temperature (b) during the growing seasons: 2003, 2004, 2005

Studies conducted 2003 through 2005 examined seeding rates ranging from 25 to 150 seeds per m2 and 20 cm and 60 cm row spacings.  Seeding rates and row spacing had no effect on seed yield during the three years (Table 2).  Between the seeding rates tested, seed yields ranged between 602 kg/ha and 710 kg/ha in 2003, 983 kg/ha and 1015 kg/ha in 2004, and 318 kg/ha and 387 kg/ha in 2005.

The two row spacings tested, i.e. 20 cm and 60 cm produced similar seed yields during the three years (Table 2).  Although the effect was non-significant, the 60 cm row spacing produced higher seed yield than the 20 cm spacing in 2003, while the 20 cm spacing produced slightly higher yields than the 60 cm spacing during 2004 and 2005.

Table 2.
Seeding rate and row spacing effects on seed yield for milk thistle

Treatment Seed yield (kg/ha)
  2003 2004 2005
Seeding rate:
25 seeds/m2 615 1015 387
50 seeds/m2 699 998 447
75 seeds/m2 747 1070 370
100 seeds/m2 682 983 364
125 seeds/m2 602 960 417
150 seeds/m2 710 1004 318
Row spacing:
20 cm 690 1028 416
60 cm 992 982 352
Analysis of variance
Source: ns ns ns
Seeding rate (R) ns ns ns
Row spacing (S) ns ns ns
R x SC.V. (%) 22.3 10.9 38.6

 

ns indicates non-significant treatment effects.

Nitrogen and Phosphorus Rate Effects:
The yield response to rates of nitrogen and phosphorus application in combination with the type of desiccant (vinegar, Reglone) and desiccation timing (30% and 60% maturity of flower head) during the 2003, 2004, and 2005 growing seasons are summarized in Tables 3, 4 and 5 respectively.

During 2003, application of 100 kg N/ha produced the lowest seed yields compared to 50 kg N/ha or no nitrogen control (Table 2.3).  In 2004 and 2005, nitrogen application in a few instances tended to produce slightly higher yield than the no nitrogen control (Tables 4 and 5).

Phosphorus application rate had no effect on seed yield of milk thistle across the three years (Tables 3, 4, and 5).

Table.3.
Nitrogen and phosphorus rate effects on milk thistle seed yield when desiccated with Reglone at two stages: 2003

Treatment Stage of desiccation
  30% mature 60% mature
Nitrogen (kg N/ha)
0 808 1121
50 782 1291
100 287 683
Phosphorus (kg P2O5/ha)
0 655 996
60 573 998
120 649 1102
Analyses of Variance
SOURCE ***(166) ***(230)
Nitrogen (N) ns ns
Phosphorus (P) ns ns
N x PC.V. (%) 32.1 26.9

*** and ns indicate significance P<0.001 level of probability and not significant respectively.  Values within parentheses are LSD estimates at 5.0% level of probability.

Table 4.
Nitrogen and phosphorus rate effects on milk thistle seed yield when desiccated with vinegar and Reglone at two different stages: 2004

Treatment Method of desiccation
  Vinegar Reglone
  30% Mature 60% Mature 30% Mature 60% Mature
Nitrogen (kg N/ha):
0 1017 1170 1308 1423
50 1079 1196 1342 1507
100 1104 1410 1289 1617
Phosphorus (kg P2O5/ha):
0 1050 1202 1289 1456
60 1043 1271 1315 1575
120 1107 1303 1335 1517
Analyses of Variance
Source ns ***(132) ns ns
Nitrogen (N) ns ns ns ns
Phosphorus (P) ns ns ns ns
N x PC.V. (%) 17.2 12.4 7.7 13.0

*** and ns indicate significance P<0.001 level of probability and not significant respectively.  Values within parentheses are LSD estimates at 5.0% level of probability.

 

Table 5.
Nitrogen and phosphorus rate effects on milk thistle seed yield when desiccated with vinegar and Reglone at two different stages: 2005

Treatment Method of desiccation
  Vinegar Reglone
  30% Mature 60% Mature 30% Mature 60% Mature
Nitrogen (kg N/ha):
0 281 364 316 212
50 304 328 317 212
100 281 353 324 288
Phosphorus (kg P2O5/ha):
0 294 397 358 248
60 281 350 301 237
120 291 298 297 227
Analyses of Variance
Source ns   ns ns
Nitrogen (N) ns ns ns ns
Phosphorus (P) ns ns ns ns
N x PC.V. (%) 21.1 32.8 20.7 48.6

*** and ns indicate significance P<0.001 level of probability and not significant respectively.  Values within parentheses are LSD estimates at 5.0% level of probability.

 

Comparison of Vinegar and Reglone as Desiccants and Timing of Desiccation:

Application of Reglone at 2.7l/ha @ 1000 l water/ha was a more effective desiccant than the commercial grade vinegar, i.e.14% acetic acid  (Figure 4).


Figure 4   Desiccated milk thistle, three days after seeding: Vinegar on the left and Reglone on the right.

Despite the large plant size and crop vigor at the time of treatment, a single application of Reglone usually produced adequate dry down.  The fact that relatively large volumes of water were used likely enhanced the efficacy of the Reglone treatment.  It was possible to combine the Reglone treated crop within 3-7 days after spraying depending the weather conditions.  Dry down was faster when the weather was warmer compared to cooler weather conditions. By contrast, vinegar was not an effective means of top-kill of milk thistle .  Even after two applications, the dry-down was insufficient after two weeks for proper machine harvesting.

On average, desiccation with Reglone produced 23% higher seed yield in 2003 and 22% higher seed yield in 2004 than using the vinegar.  This reflected seed loss during the extended dry down period required for the vinegar treatment. However in 2005, desiccation with vinegar resulted in 15% higher seed yield than desiccation with Reglone.  This is likely due greater shattering loss in the drier Reglone desiccated crop during the rain event that occurred just prior to harvest in 2005.

Delaying desiccation, from 30% mature heads to 60% mature heads, produced higher seed yields in 2003 and 2004 for both methods of desiccation.  In 2005, the early treatment increased yields in the Reglone desiccated plots – this again likely reflects differences in shattering loss during the fall storm event (Tables 3, 4 and 5).


Figure 5.  Spring seeded (left) and fall seeded (right) milk thistle.
Fall seeding was attempted for milk thistle during two years with the objective of achieving early crop establishment – thereby potentially increasing yields and seed quality.  Germination and stand establishment were extremely poor in the fall seeded crop in comparison to spring seeding (Figure 5).

Summary:
- Milk thistle shows little yield response to a wide range of in and between row seed spacings.  As seed is large and costly, low seeding rates (25 seed/m2) appear sufficient.

-  20 and 60 cm row spacings produced similar seed yields.  The wider row spacing allowed more room for early weed competition, but it also facilitated mechanical tillage to remove the weeds.

-  Fall seeding was not successful, as stand establishment was extremely poor compared to spring seeding.  The fact that volunteer milk thistles are uncommon despite the fact that the crop  shatters extensively also suggests limited overwinter survival of the seed.

- Milk thistle shows little yield response to nitrogen or phosphorus fertilizer application on a site with fairly high background levels of soil fertility.   Fertilizer applications did tend to increase vegetative growth of the crop – but this was often detrimental as it made harvesting more difficult.  Fertility levels had no impact on levels of bio-active compounds in the seed.

- The crop has to be desiccated prior to machine harvest.

- Delaying desiccation until 60% of the flowers had begun to dry down produced higher seed yields and superior seed quality (higher silymarin content).

-  Harvest timing is critical.  Inclement weather when the crop is ready for harvest (rain, wind etc.) can cause severe yield losses due to seed shattering.

-  Reglone is an efficient and cost-effective desiccant for milk thistle.

-  Commercial grade vinegar (14% acetic acid) can be used as an organic desiccant, but it is expensive, slow acting and results in increased seed loss to shattering as compared to desiccation with Reglone.

Conclusions
Over the 3 years of this study we examined the seed yield potential and level of active ingredient (% silymarin) in a range of selections obtained from commercial suppliers, breeders and gene banks.   There was very significant variability in the performance of the material available.  Some the lines failed to mature within the growing season available in Saskatchewan, while others were fully mature by late August.  There was also significant variation in seed quality as measured by the silymarin content.  There was no real association between yield potential and seed quality.   The relative performance of the lines tested was not overly consistent from year to year in terms of either their yield potential or seed quality.  This made it difficult to select for clearly outstanding lines.  In discussions with members of the Saskatchewan Herb and Spice Growers who are marketing Milk Thistle … there are at present no standards being stipulated for minimum silymarin content.  In that situation, lines with exceptional yield potential (ie; Austra Hort, Bolier or TMP 14068) would be best.  In situations where quality is also considered in developing a market price, lines that have both yield and quality would be required.  Austra Hort and  Kalyx-1 have the desired combination of yield potential with seed quality.

Wanda Wolf (pers comm 2002) estimated the break-even yield for hand harvested milk thistle at 500 kg/ha.  Yields of some of the better adapted lines in this trial were around 1500 kg/ha.  At a market price of $ 8.00 CDN/lb, a 1500 kg/ha milk thistle crop has a theoretical net return potential of $14,200/ha.

Seeding and stand establishment are straight forward for milk thistle and the crop has no problems out-competing weeds.  Fertility responses were minimal or negative (reduced stand and delayed maturity) .  Yields were not affected by either in row or between row spacings.   Milk thistle appears to be a very plastic crop, adapting to the growing space available.  As the seed is large and costly, the data suggests that a low seeding rate and wide row spacing would be most economical without compromising yield potential of the crop.  There were no significant disease problems observed and any insect pests were easily managed.

Increasing the duration of the growing season, either by early planting or delayed harvest increased yields and the concentration of active ingredients in the seed.  However, shattering of over-mature seed heads appears to be the greatest production threat for milk thistle.  The highly desirable early maturing, large seeded lines appeared to be particularly prone to shattering.  Timing and timeliness of  harvest is critical – to minimize the risk of shattering, the crop should be harvested within a few days of the seed pods reaching maturity (60% fluff) – either natural maturity or maturity triggered by frost or application of a chemical desiccant.  Once the crop reaches that stage of maturity, any delay in harvesting can result in substantial seed loss.  The desiccant Reglone provides a very quick die down of milk thistle – and consequently shattering losses are minimal.  However, Reglone is not approved for organic production.  Although application of high rates of vinegar did eventually result in desiccation of the crop, shattering losses during the extended dry down period were excessive.

Milk Thistle is traditionally hand harvested – with each seed head individually harvested as it reaches maturity.  Although this may increase yields and enhance uniformity of seed maturity, it is not a practical option in Saskatchewan.  Not only would hand harvesting be prohibitively expensive, it is doubtful that a labor pool could be identified that would be willing to work with the extremely sharp thorns on the seed heads.  Mechanical harvest using standard small plot combines appears to be a viable alternative – providing that the crop has been desiccated (by frost or chemically) prior to combining.  In two years of trials, a once over harvest with a standard small plot combine produced yields that were comparable or slightly higher than a selective hand harvest. Obtaining a yield advantage by combining was unexpected.  While harvesting the individual mature seed heads, the hand harvesters may have been bumping adjacent heads causing them to shatter out. Several of the highest yielding lines identified in this project also appeared well suited to mechanical harvesting – in that they matured uniformly, were not excessively vegetative and did not shatter prior to harvest.   Milk thistle was easily threshed, providing the tops were dead prior to harvest.   Cleaning and sorting of the seed is readily accomplished using standard equipment.  Some upgrading of quality is possible by seed sizing – but color sorting does not look to be a straight forward option.

SEEDING RATE AND DESICCATION METHOD TRIALS FOR MILK THISTLE (2007 and 2008)

Jazeem Wahab (CSIDC, Outlook, SK)
Doug Waterer (University of Saskatchewan, Saskatoon, SK)

Summary
Field trials conducted in 2007 and 2008 sought to develop appropriate agronomic management practices for the new improved lines of milk thistle recently generated by the University of Saskatchewan.  Seed yields from the new shorter stature lines were generally superior to the standard cultivar (Richters) and the seed also contained higher concentrations of the target phytochemicals.   Seed yields and seed quality increased with increasing plant populations to a peak around 100 seeds planted/m2.  This yield and quality response may reflect the more uniform flowering and seed development achieved when branching of the milk thistle plants is suppressed by growing in a dense stand.  Pre-harvest desiccation was essential to mechanical harvest of the milk thistle crop.   Diquat (reglone) and concentrated acetic acid both provided an acceptable degree of desiccation.  Diquat is faster acting and less expensive than acetic acid, but acetic acid is perceived more favorably in organic markets.   

Background 
There is substantial worldwide demand for milk thistle (Silybum marianum) seed as a herbal medicine. Both ‘conventionally’ and ‘organically’ grown milk thistle are sought by the market place.  The University of Saskatchewan has recently identified superior quality high yielding locally adapted germplasm suited for mechanical harvest under Saskatchewan growing conditions.   The Saskatchewan Herb and Spice Association has been granted exclusive rights to grow these new lines.
Previous research conducted at the Canada-Saskatchewan irrigation Diversification Centre demonstrated that there is excellent potential for mechanized production of high quality milk thistle in Saskatchewan even using the relatively unsuitable cultivars previously available (c/o Richters seed).  One of the main objectives of the present project is to develop agronomic refinements needed to maximize yields and quality of the new milk thistle lines selected at the University of Saskatchewan.  

Saskatchewan’s relatively short growing season combined with uneven maturity of milk thistle is a major challenge for production, particularly if the objective is to harvest the crop via once-over combining.  By increasing the plant population in the field it may be possible to reduce branching, thereby better synchronizing flowering, resulting in more  uniform crop maturity.    More uniform maturity should facilitate machine harvest while also reducing shattering loss and improving the overall quality of the harvested seed.

Due to its robust growth habit and indeterminent flowering pattern the milk thistle crop has to be desiccated prior to harvest.  Previous research conducted at the CSIDC and University of Saskatchewan has demonstrated that Reglone (diquat) is an effective desiccant for milk thistle.  Reglone is presently registered under regulatory review for use in Canada as a milk thistle desiccant.   However, the market preference for organic milk thistle necessitates the identification of organic desiccants for top-kill.

This project investigated agronomic practices designed to further increase yields and quality of selected lines of milk thistle in Saskatchewan, and also examined the effectiveness of vinegar as an organic desiccant.  

All the milk thistle lines utilized in this project are under exclusive licence to the  Herb and Spice & Specialty Ag Assoc.Association - therefore, all line specific information generated in this project will be of direct and exclusive benefit to Saskachewan growers of this crop

Methods  
Field trials were conducted during the summers of 2007 and 2008 at the Canada-Saskatchewan Irrigation Diversification Centre, Outlook.   Six milk thistle cultivars, (U of S-1, U of S-2, U of S-3, U of S-4, Lone Wolf, and Richters) were evaluated in 2007.  Due to the poor performance of the Lone Wolf selection in 2007, this selection was omitted from the 2008 trials.  Five seeding rates (25, 50, 100, 125, and 150 seeds/m2), two desiccants (Reglone (conventional) and Vinegar (organic) were evaluated for each cultivar separately.  Field plots were laid out on Split-plot Design with four replications.  Main-plot consisted of Desiccants and the Sub-plot consisted of seeding rates.  Individual plots (Sub-plots) were 3.7 m x 1.2 m.

The soil at the test site during the three years was a clay loam.  In the spring the land was prepared in the traditional manner to form a firm seed bed.  Plots were seeded using a double-disc press drill with seeds placed at approximately 1-2 cm depth.  Rows were spaced 60 cm apart with appropriate seeding rates as defined by the treatments.  The crop was raised under dryland growing conditions, as previous experience at CSIDC showed that irrigation of milk thistle prolonged vegetative growth and delayed maturity, resulting in considerable yield loss.  The only exception was that two light irrigations were applied after seeding to ensure germination and proper crop establishment.  One weeding during early growth stage was sufficient to raise the crop.  The crop was desiccated using Reglone and Vinegar when approximately 50% of the flower heads had matured i.e. formed of pappus on the fruit. The application rate consisted of 2.7 l/ha of Reglone applied in 1000 l water/ha during 2007 and 2008, and vinegar, (14% acetic acid at 1000 l/ha in 2007, and 60% acetic acid at 1000 l/ha in 2008).  The crop was harvested once the stalks were sufficiently dry to pass through a Wintersteiger plot combine.

Silymarin content of the harvested seed was analysed by Phytovox Inc., Edmonton, Alberta.  Pooled samples from all replicates of the lowest (25 seeds/m2) and highest (150 seeds/m2 ) seeding rates, the two desiccation treatments (Reglone and Vinegar) for each of the different milk thistle lines were analysed for Taxifolin, Silychristin, Silybin A, and Silybin B.

Trial Details

2007.  
The test plots were seeded on May 18, 2007 using a Fabro seeder.  Two light irrigations were applied after seeding to ensure germination and proper crop establishment.  The crop was otherwise grown under dryland conditions, based on previous experience with milk thistle.

The 2007 growing season was relatively warm with July being exceptionally warm and dry.  The crop received 229 mm of rain (Figure 1).  Hail storms on July 31, 2007 and August 18, 2007 caused considerable damage to flower heads, negatively affected seed filling and resulting in substantial yield losses.  The crop was desiccated on August 14 with Reglone and vinegar at appropriate concentrations.   Seed was harvested on August 30, 2007 and 31, 2007 using a Wintersteiger plot combine.  The harvested seed was dried at 35 degrees C.  The seed was cleaned and seed yield was recorded.

2008. 
Pre-plant soil analysis at the test site indicated 71 kg NO3-N/ha, 59 kg P/ha, 370 kg K/ha, and >86 SO4-S/ha at 0-30 cm depth.  No additional fertility was provided.  The test plots were seeded on May 14, 2008.  Two light irrigations were given soon after planting to ensure germination and proper crop establishment.  Subsequently, the crop was grown under dryland condition.  Flowering occurred around July 21 to July 28 and the crop was ready for desiccation three to four weeks later.  The crop was desiccated on August 22 & 23.

NB – whereas in previous trials 12-14% non-synthetic or “natural” acetic acid was used as the organic desiccant, in 2008  60% (synthetic) acetic acid was used.  The concentrated acetic acid was applied at the same rate as used in 2007 (1000 l/ha).  It should be noted that the synthetic acetic acid is not on the list of materials approved for use within “organic production systems.  

Both desiccation treatments were ready for combining within a week of treatment (August 29, 2008).

The growing season was relatively warm and received 131 mm of rain during the crop growth period.  Hail storm on August 21, 2008 caused slight damage to the crop.

Results and Discussion

Growing Season Climatic Conditions:
The 2007 and 2008 growing season temperatures were relatively similar, with July being fairly warm (Figure 1).  The crop received 229 mm of rain in 2007 and 137 mm in 2008 (Figure 2).

In 2007, hail storms on July 31 and August 18 caused considerable damage to flower heads that negatively affected seed filling resulting in substantial yield losses, but not complete crop failure.  No hail was experienced in 2008.

Seed Yield:
In 2007, the two hail events caused early death of the crop and also caused significant amounts of shattering of seed from the seed heads.  In 2008, despite lower rainfall seed yields were substantially higher than 2007.  This confirms the fact that inclement weather during the latter stages of the crop, particularly after maturity, can cause considerable crop losses.

The effects of desiccation methods and seeding rate on seed yield during 2007 and 2008 for the various milk thistle selections are summarized in Table 1 and Table 2 respectively.  The corresponding effects of treatments on average seed weight are summarized in Table 3 and Table 4 respectively.

Cultivar Response:
In 2007, the selection U of S-2 produced the highest average seed yield of 200 kg/ha, while the cultivar from Lone Wolf produced the lowest yield of 13.4 kg/ha (Table 1).  The yield ranking was U of S-2 > U of S-4 > U of S-1 > Richters > U of S-3 > Lone Wolf.

In 2008,  the average seed yield for the different selections ranged from as low as 756 kg/ha for U of S-1 up to 946 kg/ha for U of S-4 (Table 2).  The yield ranking was U of S-4 > U of S-3 > Richters > U of S-2 > U of S-1.  These seed yields are comparable to yields obtained in previous “successful” years of trials in Saskatchewan.

Seeding Rate Effects:
Seeding rate significantly affected seed yield in both 2007 and 2008 (Tables 1 and 2).  Yield responses to seeding rate were variable for the different milk thistle selections during the two years.  The mathematical relationships between seed yield and seeding rate during 2007 and 2008 for U of S-1, U of S-2, U of S-3, U of S-4, and Richters are presented in Figs 3-7 respectively.  The responses were variable for U of S-1, U of S-2, and U of S-3 between the two years. For example, the seed yield for these selections increased in a linear manner with increasing seeding rate in 2007 and assumed a quadratic relationship in 2008.  For U of S-4 and Richters, the seeding rate response was quadratic in both years.  However, it is not clear why the quadratic responses were so dissimilar over the two years for the Richters line of milk thistle (Figure 7).

The optimum seeding rate for the various cultivars can be estimated from the vertex of the yield response curve (Table 3).  The optimum seeding rate for the various milk thistle lines was somewhat variable but it appears that a seeding rate between 100 to 125 seeds per m2 tended to produce the highest seed yields

The two desiccation methods used, i.e. Reglone or Vinegar, provided an equivalent degree of crop desiccation and had no effect on seed yield in 2007 and 2008 (Tables 1 and 2).  In 2007, the crop had been severely damaged by hail prior to application of either top-killing treatment.  This hail damage may have rendered the crop susceptible to even the relatively mild effects of desiccating with relatively dilute (12%) acetic acid.

In previous trials ( Milk Thistle Agronomy) the Reglone treatment had consistently produced higher seed yields than the organic desiccant.  This yield advantage was attributed to the much more rapid desiccation achieved with Reglone – as a rapid drydown would reduce the time during which the mature seed could shatter out.  In the 2008 trial the drydown achieved with the “organic” desiccant was as quick and thorough as that achieved with the Reglone.  This may be atrributed to (i) the change over to more adapted, earlier maturing milk thistle lines, (ii) use of production practices conducive to acclerated crop maturity (ie; reduced fertility, minimal irrigation, high seeding rates) and c) the use of much more concentrated acetic acid (60%) in 2008 versus 12% acetic acid in 2007.

Average Seed Weight:
Seeding rates and desiccation methods had no effect on seed weight for any of the milk thistle lines tested in 2007 and 2008.  Average seed weight (g/1000 seed) for the different selections varied between 16.5 to 18.4 g in 2007 (Table 4) and from 14.7 to 22.2 g in 2008 (Table 5).  U of S 3 produced the smallest seed (14.5 g/1000 seed); U of S 2, U of S 4, and Richters produced larger seeds (20.5 – 22.2 g/1000 seed), and U of S 1 produced intermediate seed (16.5 g/1000 seed).

Quality (Flavonoids):
The flavonoid profile for the various milk thistle selections was influenced by seeding rates and desiccation methods.  These results are summarized in Table 6 (individual values) and Table 7 (treatment averages).

This preliminary analysis showed the following trends with respect to the effects of treatments on flavonoid for the various milk thistle selections:

-  seeds from U of S 2 at 150 seeds/m2 desiccated with Reglone and U of S 1 at 25 seeds/m2 also desiccated with Reglone contained the highest and lowest amounts of all the flavonoids respectively.  The total flavonoid levels ranged between 11.48 and 21.66 µg/g (i.e. 1.2% and 2.2%) of seed dry weight.

- U of S 2 contained the highest amounts of Taxifolin, Silychristin, Silybin A, Silybin B, and Total flavonoids relative to the other selections.  In previous trials U of S 2 had consistently produced superior levels of the various bioactive molecules (see Milk Thistle Agronomy).

- seeds harvested from plots with the highest plant population (150 seeds/m2) had higher flavonoid levels than seed from plots with the lowest plant population (25 seeds/m2).   This may reflect the earlier and more uniform seed maturity achieved with a higher plant population.   The flavonoid content of milk thistle seed is know to increase with seed maturity, although it peaks and stabilizes well before the seed is ready to drop from the seed head.   The relative amounts of the different flavonoids are also known to change with seed maturity, however these changes are also strongly influenced by genotype, growing conditions and method of extraction and analysis.    There is still some considerable debate within the industry as to which of the flavonoid components is the best indicator of “potency” of the milk thistle extracts.  Until this debate is settled, the overall silymarin content is considered to be the “default” indicator of quality.

- there was no difference in flavonoid levels between the two desiccation treatment (Reglone or Vinegar).
Conclusions

The objective of the agronomy trials conducted in this project was to develop appropriate management practices for the improved lines of milk thistle recently generated by the University of Saskatchewan.  The new milk thistle lines produced a vigorous stand of relatively short stature evenly maturing plants that were better suited to mechanical harvest than the line previously used as the industry standard (c/o Richters).  Seed yields from the new lines were generally superior to the standard cultivar, although the performance of all lines varied from year to year.  The nature of this genoype X environment  interaction needs to be further explorer in order to increase confidence in the performance of a specific line(s).   The seed harvested from the new lines also contained higher concentrations of the target phytochemicals than the standard.    The combination of improved yields and superior seed quality indicates that the new lines created by the University of Saskatchewan are meeting expectations.   In particular, the line UofS-2 has emerged as providing a combination of superior seed yields and seed quality in a vigorous but readily managed short stature, fast maturing plant phenotype.

Seed yields and seed quality increased with increasing plant populations to a peak at around 100 seed/m2.   This yield and quality response to plant population may reflect the more uniform flowering and seed development achieved when branching of the milk thistle plants is suppressed by growing in a dense stand.    The dense stand also reduced weed competition early in crop development.  While the heavy seeding rate would increase seeding costs, the corresponding yield and quality advantage is clearly sufficient to offset this cost.

This study, as well as previous research showed that there was a significant year X genotype interaction for both the total concentration and the relative amount of different bioactive flavonoids found in the milk thistle seed extract.   It also appeared that the flavonoid content and composition could be influenced by plant populations but was not influenced by the method of desiccation.   These interactions complicate the process of identifying and adopting genotypes and production practices that optimize product quality.  This process is further complicated by the fact that the industry has not been able to identify which of the flavonoid components is the best indicator of “potency” of the milk thistle extracts.  Until this debate is settled, the overall silymarin content should likely be considered to be the “default” indicator of quality.   This project has identified genotypes and production practices designed to maximize both yields and product quality.

While the new short stature, even maturing milk thistle lines are clearly better suited to mechanical harvesting than the previous standard lines, pre-harvest desiccation is still essential to efficient mechanical harvest of the milk thistle crop.   Diquat (reglone) and concentrated acetic acid both provided an acceptable degree of desiccation.   It should be noted that thorough coverage is essential for effective kill down of a vigorous milk thistle crop – and this coverage hinges on the application of very large volumes of the desiccant (1000L/a).   Growers must be cautioned against trying to scrimp on the desiccant volume – otherwise the top-kill process will be compromised.  Of the two desiccants tested, Diquat is more readily available, faster acting and much affordable.   Registration of reglone for use in milk thistle is undergoing regulatory review in Canada.   The main reason that a grower would consider use of acetic acid is demand in the organic marketplace.    This raises a potential problem.   In previous research the naturally sourced 12% acetic acid provided a slow and incomplete degree of top kill even at very high rates of application.   While this product provided an acceptable degree of top kill in the 2007 crop it should be noted that this crop had been weakened by a severe hail event and was therefore relatively easy to kill down.   For this reason it was decided to test the efficacy of a much more concentrated form (60%) of acetic acid in 2008.   This formulation provided the very fast and thorough desiccation required for consistent machine harvest of milk thistle.  However, it should be noted that this form of acetic acid is derived from a synthetic process that uses petrochemicals as starting materials.   This effectively disqualifies this product from approval for use in “certified organic” production.    In discussions with the SHSA, the issue of which product a milk thistle crop is desiccated with has not raised any particular concern.  The market is not presently requiring organic certification and is not paying a premium for organically grown milk thistle.   This would suggest that in the present market situation desiccation with either Reglone or concentrated acetic acid are viable options.


Table 1: Seeding rate and desiccant effects on yield for milk thistle line selections: 2007 trial
Treatment Seed Yield (kg.ha)
  U of S-1 U of S-2 U of S-3 U of S-4 Lone Wolf Richters
Desiccant
Reglone 158.1 184.9 74.6 183.0 14.4 107.7
Vinegar 162.7 214.8 68.9 201.4 124 118.7
Seeding Rate
25 seeds m/2 110.1 99.4 47.8 138.9 3.5 58.8
50 seeds m/2 158.4 144.3 70.3 194.3 7.6 100.1
125 seeds/m2 172.3 249.8 91.8 171.6 20.6 132.6
150 seeds/m2 203.7 286.5 75.7 211.3 17.3 143.9
ANOVA
SOURCE
Desiccant ns ns ns ns ns ns
Seeding Rate **(64.9) ***(69.8) *(38.4) ns **(12.4) ***(51.4)
Desiccant x Seeding Rate ns ns ns ns ns ns
C.V. (%) 27.7 24.0 36.6 40.7 64.5 31.1
***, **,* and ns indicate significance at P<0.01, 0.05 levels of significance and not significant respectively. Values within parentheses are LSD estimates at 5.0% level of significance.

 

Table2;
seeding rate and desiccant effects on seed yield for milk thistle line selections: 2008 trial

Treatment Seed yield (kg/ha)
Desiccant
Reglone 738.8 789.0 898.2 1029.4 871.7
Vinegar 773.4 801.1 841.2 862.4 849.8
Seeding Rate
25 seeds/m2 651.9 627.8 684.3 741.5 846.4
50 855.3 722.4 860.4 915.2 835.8
100 756.7 853.4 935.2 1097.5 833.1
125 813.0 920.8 917.7 937.5 858.7
Analyses of Variance
Source
Desiccant (d) ns ns ns ns ns
Seeding rate (R) *(137.4) ***(117.7) **(120.1) *(240.9) ns
DxR ns ns ns ns ns
C.V. (%) 17.6 14.3 13.4 24.7 22.5
***, **, * and ns indicate significance at P<0.001, 0.005 levels of probability and not significant respectively. Values within parenthesis are LSD estimates 5.0% level of significance,

 

Table 3.
Summary of mathematical ralationships between seeding rate and seed yield of milk thistle line selections: 2007 and 2008

Selection Year Response R2 Vertex
U of S-1 2007 Linear 0.82 -
U of S-1 2008 Quadratic 0.52 89
U of S-2 2007 Linear 0.99 -
U of S-2 2008 Quadratic 0.96 125
U of S-3 2007 Linear 0.62 -
U of S-3 2008 Quadratic 0.92 121
U of S-4 2007 Quadratic 0.55 102
U of S-4 2008 Quadratic 0.78 113
Richter's 2007 Quadratic 0.97 139
Richter's 2008 Quadratic 0.96 -

 

Table 4.
Seedling rate and desiccant effects on average seed in weight for milk thistle line selections

Treatment 1000 seed weight (g)
  U of S-1 U of S-2 U of S-3 U of S-4 Lone Wolf Richters
Reglone 18.0 17.5 18.5 16.0 16.7 17.0
Vinegar 18.2 18.2 18.2 17.6 16.8 16.0
Seeding rate
25 seeds/m2 18.9 16.6 17.7 18.2 15.1 14.6
50 seeds/m2 17.8 17.4 18.4 16.5 18.7 16.6
100 seeds/m2 18.0 17.6 18.0 17.4 17.8 16.8
125 seed/m2 18.0 19.0 18.6 14.4 16.0 17.4
150 seeds/m2 17.9 18.7 19.0 17.6 16.2 17.2
Anova
Source            
Desiccant ns ns ns ns ns ns
Seeding rate ns ns ns ns ns *(2.7)
Desiccant X Seeding rate ns ns ns ns ns ns
C.V. (%) 6.7 13.7 9.9 18.6 24.0 11.2
* and ns significance at P<0.05 level of significance and not significant respectively. Value within parenthesis is LSD estimate at 5.0% level of significance

 

Table 5
Desiccation and seeding rate effects on average seed weight of milk thistle line selections: 2008 trials

Treatment 1000 seed weight (g)
  U of S-1 U of S-2 U of S-3 U of S-4 Richter's
Desiccant
Reglone 16.7 22.0 14.7 20.5 21.1
Vinegar 16.2 22.4 14.3 20.4 21.0
Seeding rate:
25 seeds/m2 16.7 22.4 13.5 20.4 21.5
50 seeds/m2 16.8 21.6 15.0 21.1 20.9
100 seeds/m2 16.1 22.1 14.7 20.5 21.2
125 seeds/m2 16.0 22.6 14.2 20.0 20.9
150 seeds/m2 16.7 22.4 15.1 20.2 20.7
Analyses of variance
Source          
Desiccant (D) ns ns ns ns ns
Seeding rate (R) ns ns ns ns ns
DxR ns ns ns ns ns
C.V. (%) 10.9 5.1 2.0 4.8 8.0
ns indicate non significant treatment effects.

 

Table 6.
Effects of seedinf rate and desiccation on seed flavonoid content in milk thistle selections

Selection Dessication Seeding rate
(seeds/m2)
ug per mg dry weight
      Taxifolin Silchristin Silybin A Silybin B Total
U of S-1 Reglone 25 1.371 3.896 2.661 3.549 11.48
U of S-1 Reglone 150 1.747 4.782 3.407 4.484 14.42
U of S-1 Vinegar 25 2.102 5.066 3.525 4.669 15.36
U of S-1 Vinegar 150 2.115 5.747 4.097 5.367 17.33
U of S-2 Reglone 25 2.496 6.078 4.551 5.834 18.96
U of S-2 Reglone 150 2.998 6.984 5.081 6.593 21.66
U of S-2 Vinegar 25 2.514 5.732 4.204 5.484 17.63
U of S-2 Vinegar 150 2.746 6.451 4.769 6.189 20.15
U of S-3 Reglone 25 1.562 4.319 2.987 3.921 12.79
U of S-3 Reglone 150 1.747 4.692 3.222 4.369 14.03
U of S-3 Vinegar 25 1.591 4.668 3.279 4.304 13.84
U of S-3 Vinegar 150 1.863 5.366 3.811 4.962 16.00
U of S-4 Reglone 25 1.405 5.354 3.852 5.123 15.73
U of S-4 Reglone 150 1.454 4.98 3.548 4.644 14.63
U of S-4 Vinegar 25 1.199 4.378 3.099 4.156 12.83
U of S-4 Vinegar 150 1.586 5.476 3.892 5.182 16.14
Richter's Reglone 25 1.876 4.895 3.423 4.528 14.72
Richter's Reglone 150 2.204 5.176 3.557 4.703 15.64
Richter's Vinegar 25 1.178 4.558 3.234 4.26 13.77
Richter's Vinegar 150 1.956 4.350 2.960 3.948 13.21

 

Table 7.
Treatment effects on average flavenoid levels in seed of milk thistle line selections

Treatment ug per mg dry weight
  Taxifolin Silychristin Silybin A Silybin B Total
Selection
U of S-1 1.834 4.873 3.423 4.517 14.65
U of S-2 2.689 6.311 4.651 6.025 19.6
U of S-3 1.691 4.761 3.325 4.389 14.17
U of S-4 14.11 5.047 3.598 4.776 14.83
Richter's 1.939 4.745 3.294 4.36 14.34
Seeding Rate
25 seeds/m2 1.783 4.894 3.482 4.583 14.71
150 seeds/m2 2.042 5.4 3.834 5.044 16.32
Desiccant
Reglone 1.886 5.116 3.629 4.775 15.41
Vinegar 1.939 5.179 3.687 4.852 15.63

 

MILK THISTLE VALUE ADDED

Value-added Products From Milk Thistle “Waste”

At present, only a small component of the milk thistle plant (silymarin from the seed) has market value.  The seed contains significant qualities of oil – yet this oil is presently regarded as a waste product.  The plant also produces a huge amount of biomass – another waste product unless alternate uses are discovered.  Potential uses of the waste biomass are animal fodder or alternatively a starting point for biofuel – especially as the stems and leaves are rich in high energy latex compounds.  These latex compounds may also have value as a feedstock in the manufacture of industrial materials … like rubber and latex.  At present, there are no crops being grown for latex in Canada – but there is demand for latex as the starting point in the manufacture of rubber and other materials.  The potential to extract products of potential use as fuel, food, fodder and medicine from a single easy to grow plant would appear to make milk thistle an ideal model for of a future bio-based economy.

This project sought to identify potential value in the “waste products” generated by the milk thistle crop.

1.0       Recovery of Value-added Products from Milk Thistle

1.1.      Seed oil

The research literature indicates that milk thistle seed contains 12-26% fixed oils, depending on the genotype and the growing conditions.  At present, these oils represent a waste product which is discarded following silymarin extraction from the seed.

This project evaluated the oil concentration, composition, total yield and potential value of the oils recovered from the newly developed milk thistle lines being commercialized by the SHSA.

Experimental design
Seed source – SHSA lines generated in agronomy trials
Variables examined
- Total oil content
- Oil composition
- Impact of agronomic variables and site/growing season on oil yields

Milk Thistle seed samples covering the previously outlined range of variables (cultivar, site of production, year of production) were sent to Dr. Martin Reaney of the Dept. of Applied Microbiology and Food at the University of Saskatchewan for analysis.

Oil was extracted from small samples of milk thistle seed using a Goldfisch extraction apparatus and the oil composition was than tested via NMR.   A cold press was used to extract the oil from larger (1 kg) samples of a limited number of seed lots.

Averaged over the cultivars tested, seed oil content was lower in 2004 (20.2%), than in 2005 (25.2%), or 2006 (24.9%)(Table 1.1).  The low oil content in 2004 likely reflects the immature stage of the crop when hit by an early killing frost.  The 2005 and 2006 growing seasons were much more favorable, leading to higher yields of more mature seed with a higher seed oil content.  Differences in seed coil content between cultivars were not consistent from year to year.  Oil yields from cold press extraction were comparable to the yields obtained using the Goldfisch apparatus (Table 1.1).

Table 2.1.   Seed oil content for various milk thistle lines in 2004-2006.

NMR analysis of the oils showed no significant differences among the samples (data not shown). The fatty acid profile of all samples was essentially similar (Table 1.2).

 

Table 1.2  Fatty acid profile of the oil extracted from the seed harvested from several cultivars of milk thistle in 2004-2006
Fig 1.1.  NMR analysis of the oil profile typical of Milk Thistle.

The seed oil content and fatty acid profile of milk thistle oil, along with several of the most commercially important oils used in food production are summarized in Table 1.3.

 

Table 1.3
Seed oil content (%) and fatty acid profile (%) of oil extracted from Milk Thistle seed and several other commercially important oil crops.

The fatty acid profile of the oil extracted from milk thistle was comparable to the oil extracted from corn, sunflower and soybean.  This suggests that the oil would likely have acceptable performance characteristics for use in food processing – assuming that it does not have any unusual/undesirable flavour characteristics.  However, the fact that the oil composition is comparable to mainstream commercial products like corn and soybean oil means that it will be in direct market competition with these low cost alternatives.

Based on the seed yields and seed oil content seen in this study, oil production/unit area for milk thistle is well below that of the key oil producing crops (Table 1.4).  The low oil yield/unit area, coupled with the fact the fatty acid profile is not unique suggests that growing milk thistle as an oil crop is not an attractive option.  However the results do suggest the potential to use the milk thistle oil that is left over from silymarin extraction as another profit stream – much like the wine industry is capturing value (grape seed oil) from the residues left from their main processing objective.

Table 1.4.        Oil production/unit area for milk thistle and other oil producing crops

2.1 Identification/Recovery of other useful by-products from milk thistle

Milk thistle gets its name from the “milky” exudate that oozes from any cuts to the leaves, stems or roots.  The chemical composition of the milk thistle exudate is not known – but closely related species have been identified as sources of latex.  The fact that milk thistle produces a large biomass with minimal management or inputs, suggests its potential efficiency as a ‘bio-factory’.  As the vast majority of the milk thistle plant is, at present, effectively a waste product – any alternate uses for this waste material would represent another potential profit source.

Test material grown in 2006 and 2007 by the University of Saskatchewan was made available to Dr. John Balsevitch, G. Bishop and L. Deibert – who are natural products biochemists employed by PBI/NRC, Saskatoon.

Leaves of milk thistle were harvested just past full flowering.  The leaves were allowed to air dry and processed via grinding and followed by extraction with 60% methanol (aq) followed by 100% methanol.  A 50 g sample of dried leaf material yielded after evaporation of the combined methanolic extracts about 15 g of residue.

The combined methanolic extract was analyzed by HPLC-MS and observed to contain two main flavonoid glycuronides tentatively identified as apigenin and luteolin 7-O-glucuronides (Fig 2.2.1) based on the spectral data and a published report of similar compounds being observed in the flowers.1

In the samples examined the putative “apigenin” was the major flavonoid (ca. 7:3 ratio).

Processing of the extract via partitioning between ethyl acetate/water and butanol/ water afforded three fractions – ethyl acetate soluble (1.1 g), butanol soluble (1.8 g) and water soluble (12 g).  The ethyl acetate fraction was composed of lipophilic materials including pigments and triterpenes, the butanol fraction was composed largely of the flavonoid glucuronides, while the aqueous fraction contained polar materials believed to be unidentified salts and amino acids/proteins.  Samples of relatively pure flavonoids could be obtained by further processing the flavonoid containing extracts on a reverse phase column using gradient water-methanol elution.

Flavonoids in Milk Thistle
In general, flavonoids are widespread in the plant kingdom but glycuronide derivatives are not the most common representatives – glycosides and aglycones are much more common.  Flavonoids are mostly anti-oxidants, considered to possess nutraceutical qualities, and by and large, generally recognized as safe  (GRAS status).   Apigenin glucuronides have previously been isolated from alfalfa leaves 2 and galacturonides identified in milk thistle flowers1.

Searching the patent literature afforded two3,4  patents which describe utilization of flavonoid glycuronides.  In one they were observed to promote improved solubility and uptake of various drugs; in the other they were used to prolong the life of natural anthocyanin pigments used as coloring agents.

Fig. 2.2.1. Putative structures of major (1) and minor (2) flavonoids observed in air-dried mature Silybum marianum leaves obtained from field plants, University of Saskachewan 2006/07.


2.2.2    Non-Flavonoid Components

Examination of the hplc-ms data (see Appendix 2.1) led to the tentative identification of triterpenes having a molecular weight of 456, possibly ursolic and oleanolic acids.  Previously reported5 oxygenated triterpenes were not observed.  The leaves are also considered a rich source of amino acids (both free and in proteins, (14%) as well as magnesium and potassium (source: http://www.ars-grin.gov/cgi-bin/duke/farmacy2.pl) and are considered edible on removal of the spiny thorns.  The bulk of the extract was the water soluble material which, based on the above report, was likely a mixture of salts, amino acids, proteins, and possibly polysaccharides.

 

2.2.3    Bioassays/Reported Activities

The major flavonoid (MW 446) was evaluated in hemolysis and apoptosis assays.  It was found not to be hemolytic in sheep red blood cells (HD50 >100 μM) and did not cause any increase in caspase 3/7 activity or mitochondrial pertabation in PC-3 human prostate cancer cells at 50 μM (Dual Sensor MitoCasp assay using flow cytometry).  Both of these assays suggest that the flavonoids have low cytotoxicity.  This is further corroborated by historical usage of the leaves as a tea:

“The flowers and leaves of the milk thistle plant can also be used in an infusion. A milk thistle hot tea is used to stimulate milk production in nursing mothers, as well as to treat digestive problems.” (http://www.disability-resource.com/medical-health/alternative-medicine/herbal/antioxidant-rich-milk-thistle-herbal-remedy.php).  The above results in combination with reports of historical usage suggest low toxicity of the contained flavonoids.
Leaf extracts have been shown to extert anti-inflammatory properties in rats.7

 

2.2.4.   Conclusion: Potential Utility of Aerial Parts of Plant

Since the aerial part of the plant represents quite a large amount of biomass, suitable processing could lead to quantities of triterpenes (ursolic and oleanolic acids), potentially useful in cosmetic formulations (based on published anti-inflammatory properties), apigenin/luteolin glucuronides potentially useful in nutra- and /or pharmaceutical applications as outlined in the cited patents, and protein/amino acids potentially useful for food/feed applications.

For commercial development, nutraceutical claims for the flavonoids would have to be established and identification of salts and amino acids present in the aqueous fraction would be required.

Finally, development of milk thistle without thorns would be interesting – perhaps the plant could then be developed as a nutritious salad crop for human use or as a forage crop.  Allergy considerations may temper potential food uses.

Appendix:  2.1  HPLC-MS Data for Milk Thistle Leaf Extracts

Fig. 2.2.1.  HPLC profiles of extract obtained from air-dried mature leaves of milk thistle.  Top: uv detection @ 280-340 nm; Middle: uv detection @ 200-400nm; Bottom: mass detection (m/e 100 – 1900).

Fig 2.2.2.  UV spectrum of compound 1 and 2.

Fig. 2.2.3.  Mass spectrum (electrospray, positive ion) of compound 1, showing loss of galacturonate (or glucuronate).

 


Fig. 2.2.4.  Extracted ion  (m/z 456) mass chromatogram of MT leaf extract.  Peaks tentatively assigned as oleanolic and ursolic acids

GRADING TO IMPROVE MILK THISTLE QUALITY
It has been established that the quality of milk thistle increases as the seed matures. Therefore the quality of a given seed lot can be improved if the seed can be sorted for maturity. Immature milk thistle seeds are small, thin and quite light in color by comparison with mature seed. Milk thistle seed cab be easily sorted with readily available equipment based on size and density. This does an acceptable job of sorting out the very immature seeds … but cannot achieve any further grade improvement. It is possible however, to further sort In 2003, we used a color sorter to grade a seed lot (Richter’s) and then examined based on seed color.

thesilymarin content of the resulting grade categories. There was some tendency for the silymarin content to increase with increasingly dark seed color. In 2004, we ranked the various cultivars tested in the germplasm evaluation trial for seed coat color (darkest = rank of 1) and silymarin content (highest = rank of 1)(Figure 1.7). We found that there was little direct association between seed coat color and silymarin content. We also ranked various lots of the cv. Richter’s grown in differing trials (* in Fig. 1.7). Again, there is little in the way of a relationship between coat color and silymarin content. This suggests limited potential to use color sorting technology to further .improve sorting for quality in milk thistle

/you have reached the end of Objective 3 - Click here to return to the top of the page

.

 

Objective 4. Double Haploid Trials.

download the .pdf

Medicinal and aromatic plants share common breeding objectives with most traditional crops – the industry is striving for increased yields and superior quality. Traditional breeding programs are costly and take years to achieve results. This is particularly true in self pollinating species and in situations where extensive backcrossing is required to introduce a new trait into a crops with complex quality and yield expectations.
In the double haploid (DH) technique used by PBI/NRC (Alison Ferrie) immature pollen grains (haploids) are exposed to treatments that result in doubling of the existing genetic material – resulting in homozygous, true breeding material in a single generation. This technique has been used to reduce the length of the crop improvement cycle of crops like canola and wheat by several years. Dr. Ferrie has also applied the DH technique to a wide range of medicinal and nutraceutical plant species.
The objective of this aspect of the program was to evaluate the field performance of the double haploids of various herb and spice crops that have been created by Dr. Ferrie

 

CONTENTS BELOW INCLUDE:

- "2003 Trials"

- "2004 Trials"

- "2006 to 2008 - Dill Trials"

-"2006 to 2008 Caraway Trials"

- "Field Evaluation of Doubled Haploid Plants in the Apiaceae: Dill, Caraway and Fennel"

2003 Trials
In 2003, seed of double haploids of dill and fennel were obtained from Dr. Alison Ferrie of PBI/NRC. Due to limited seed supplies, the double haploid lines were grown from transplants. Seed was pre-germinated on petri dishes and then transferred to transplant flats held in the Dept of Plant Sciences greenhouses. The transplants were moved into the field when they were 6 weeks old.

The field plots were located at the Department of Plant Sciences, Horticultural Crops Research Station in Saskatoon. The features of the site were previously described in the Cumin Germplasm Evaluation Section.

The seedlings were planted out in early June. Rows were spaced 0.5 m apart, with 15 cm between seedlings within a row. The number of plants of each line varied according to the % of the seeds planted that actually germinated. Weeds were controlled by hand removal. The plots were watered as needed. No significant problems with insect pests or disease were observed. Plant heights were recorded at several points during the growing season. The crops were harvested in late September, after the first killing frosts but before significant shattering occurred.

Results
Germination percentages varied greatly between crops and DH lines. In general, the parental lines of dill had a higher germination % than the DH, but some of the DH lines of dill showed excellent seed viability. A few of the DH lines of dill completely failed to germinate. The parental material in the fennel trial had excellent seed viability. This was matched by about 50% of the DH lines; others had a lower % germination and a few DH lines of fennel failed to germinate. .

Plants heights also varied between DH lines and the parental strains. Eight of the nine DH lines of dill tested were substantially shorter than the parental line. However, one line (DH-1) was significantly taller than the parental line. The DH lines of dill were also much more uniform in height from plant to plant than the parental lines. Heights of the DH fennel were either very similar to the parental line or were nearly double the norm. There were too few DH anise plants to effectively judge plant heights.

Seed yields for a number of the DH lines of dill were substantially greater than yields from the parental lines. The three most productive DH lines averaged 61% higher yields than the parental line. There was no apparent relationship between the height of the DH plants and their productivity.

The quantity and quality of the oil produced by the DH dill lines was evaluated by solvent extraction following by gas chromatography.

 

The essential oil content of many of the DH lines was lower than the parental line (3.3%)(Table 4.1). However, DH lines DH-1 and DH-12 had substantially higher essential oil content than the parental line. DH-12 also produced higher yields than the parental line. The oil extracted from dill should contain > 50% carvone. All of the DH lines except for DH-1 met this quality criterion.

Table 4.1. Essential oil content and quality for double haploid lines of dill – 2003.

Sample % Essential Oil % Carvone
Mammoth (Parental) 3.3 52.8
DH-1 3.9 44.2
DH-4 1.3 57.5
DH-7 2.6 51.8
DH-12 3.8 51.1
DH-19 1.6 51.8
DH-24 1.8 55.7
DH-35 1.9 52.5
DH-42 2.1 51.9
DH-45 1.5 51.0

 

2004 Trials
In 2004, the main trial involved direct seeding the DH dill lines tested the previous season using transplants. The trials were run at the Department of Plant Sciences, Horticultural Crops Research Station in Saskatoon. This features of the site were previously described.

The site was prepared by rotovating prior to seeding. The plts were seeded on May 14 using a pushtype precision small plot seeder. The trial was seeded with 5 cm between seeds within a row and 50 cm between rows. Each plot of each line consisted of four 6 m long rows. Weeds were controlled by hand removal. The plots were watered as needed. No significant problems with insects pests or disease were observed. Plant heights were recorded at intervals during the growing season. The crops were harvested on October 13, after the first killing frosts but before significant shattering had occurred. The plots were harvested using a Wintersteiger Nursery Master Elite small plot combine. The harvested material was air dried at 40 C for 5 d, further threshed and then cleaned using a dockage tester. Seed yields and 1000 seed weights were determined at this point.

Results

Direct seeded dill trial.
As the trial was seeded quite heavily, all lines produced a complete stand. Line DH-12 emerged exceptionally quickly, established a very strong stand and was the first line to begin flowering. At the time of evaluation in late July, DH-12 was the tallest line and was furthest advanced in terms of flower development (Figure and Table 4.2). Line DH-12 also showed great uniformity of plant height at that time. However, by the final harvest line DH-12 was the shortest line but its seed yields exceeded all others, including the parental line by a factor of 300 % (Figure and Table 4.2).

The precocious characteristics exhibited by DH-12 appeared highly beneficial in the relatively cool and short 2004 growing season. None of the lines were actually very mature by the time of harvest, but line DH-12 was more advanced than the others. Seed size for DH-12 was comparable to the parental line. None of the other DH lines produced seed yields that exceed the parental line. Although line DH-7 had yielded well in the 2003 trial, it performed poorly in 2004. Line DH-1 was slowest to develop but by the en d of the growing season it was substantially taller than the parental line. Seed yields from DH-1 were poor as it is slow to develop, however the seeds that it did produce were exceptionally large (Table 4.2). The unusually large size of the DH-1 plants may make it desirable for production of dill oil which can be extracted from both the seed and the herbage.

Table 4.2
Plant heights, yields and seed size for various lines of direct seeded double haploid dill in 2004.

  JULY 26 FINAL HARVEST
  Height
(cm)
Height
cv %)
Stage Height
(cm)
Height
(cv %)
Seed Yield
(kg.ha)
1000
seed wt
(g)
DH 1 51e 22 0 175 a 29 76 c 3.55 a
DH 7 64 d 21 1.5 144 cd 29 54 c 2.67 ab
DH 12 88 a 5 5 125 e 29 1205 a 1.94 b
DH 24 72 c 9 2 146 c 29 230 bc 2.12 b
DH 35 82 b 12 2 136 cd 30 481 b 1.70 b
DH 45 72 c 14 1.5 135 de 30 286 bc 2.09 b
Parental 78 b 9 2 159 b 30 344 bc 3.02 b
Values within columns followed by the same letter are not significantly different p = 0.05

 

Table 4.4. Essential oil content and quality for double haploid lines of dill – 2004.

 

Sample % of Essential Oil % Carvone % Limonene Oil yield (kg/ha)
Mammoth (Parental) 2.91 57.4 38.5 1001
DH 1 2.38 55.0 40.7 180
DH 7 2.48 51.8 40.0 133
DH 12 2.63 53.4 43.9 3169
DH 24 2.38 51.2 44.1 547
DH 35 2.12 49.6 45.8 1019
DH 45 2.45 59.0 46.4 700
DH 47*** 2.51 46.8 46.3 8117
DH 53*** 3.11 52.9 43.3 7501
*** = transplanted single row plots
On average, essential oil content in 2004 was somewhat lower than in 2003; this likely reflects differences in the growing season. In 2003, several of the DH lines had a higher essential oil content than the parental line – but in 2004 the parental line had the highest oil content. Most lines met quality standards for % carvone (50%). Oil yield (kg/ha) for line DH 12 was 300% of the parental line. This reflects both its high seed yield and relatively high oil content. Oil yields for lines DH 47 and DH 53 were both exceptionally high – but as previously mentioned, this may reflect the advantage provided by transplanting and wide row spacings.

Fennel

As in the 2003 trial, the germination % of the DH fennel lines was quite variable. Many lines appeared to be non-viable while a couple showed excellent vigor.

Plant heights for the DH fennel lines in 2004 did not show the extreme line to line variability that was observed in the 2003 trial. Due to the short, cool growing season, no mature seed was harvested from this trial.

2006 - DH Dill Trials
-pdf

In 2006, the main testing trial utilized the DH lines of dill generated by PBI/NRC that had shown the most promise in small scale screening trials conducted in 2004 and 2005. Lines were selected and multiplied for further testing if they showed promise in any of the following areas;

a) plant vigor, uniformity of height, uniformity of maturity
b) seed yield
c) oil content of the seed
d) chemical composition of the seed oil.

Key objectives were to identify lines that retained the seedling vigor of the parental lines, but were shorter and more uniform in stature. Improved uniformity of seed set was a highly desirable characteristic. The presently available commercial dill lines have a very indeterminate growth habit resulting in uneven seed maturity. With highly indeterminate seed set it is inevitable that a portion of the seed is lost to shattering before the crop is ready to harvest … while a portion of the seed is also still immature at harvest. Immature dill seed does not have the flavor profile required by the industry.

All field trials were conducted at the University of Saskatchewan Horticulture Research Facility in Saskatoon (2909-14th Street East). This field site features a Sutherland Series heavy clay soil. This soil type is less than ideal for establishment of small seeded crops – but once the crops get established, its moisture and nutrient retention characteristics result in excellent yields. The test site had been in summer fallow in 2005.

The field was rotovated prior to planting to provide as uniform a seed bed as possible. No supplemental fertilizers were required. The trial was seeded on May 23. This is a later planting date than most commercial growers would use – but it reflects the fact that the clay soil at the test site is very slow to dry and warm in the spring. The crop was seeded with a push-type small plot seeder. Each line was planted out in 3 rows (0.5 m apart) with each row being 8 m long. Each plot was separated by a 1 m buffer. The trial was laid out in a randomized complete block design with 3 replicates.

Most of the seed used in the 2006 trials was generated in 2004 and 2005 in small plot observation trials conducted by the U of S. Because of adverse growing conditions in 2004 and problems with herbicide injury in 2005, the viability of the seed was relatively low and also varied between lines. This was confirmed in germination tests conducted prior to seeding the 2006 crop (Table 2006-1). To compensate for the low seed viability, we opted to use an exceptionally high seeding rate – roughly 4X normal. This resulted in a very dense stand in the lines with the highest germination % , but only a spotty stand in the lines with the poorest germination %. Like canola, dill is an indeterminate crop. Dill plants are capable of altering their branching habit to fully utilize the available growing space. Consequently, once the plants reached maturity, the plots for the various lines were relatively uniform in terms of canopy density.

Warm weather coupled with timely rainfall in early June resulted in rapid and uniform crop emergence but also produced near-ideal conditions for the germination of weeds. The herbicide linuron (Lorox) was applied (200 g a.i./a) once the crop reached the 3 true leaf stage. The herbicide was lightly watered in 4 days after application. Linuron is cleared for use on a wide range of Umbelliferous crops (ie; coriander and carrot) – but it is not presently registered for use on dill. Rick Holm (weed scientist at the U of S) has previously tested linuron on dill and found it to be quite crop safe. However, when this product was used on the 2005 seed increase of the DH dill, it caused significant crop damage (yellowing and some seedling death). This problem was initially attributed to adverse conditions following application of the herbicide. However, similar problems occurred again in 2006 – and in that case conditions at and following application of the linuron were near-ideal. The degree of damage was similar in both the parental lines and the double haploids. Relatively few plants actually died as a result of the herbicide treatment in 2006, but growth of the crop was set back by at least 3 weeks. The linuron provided excellent season-long weed control in this trial.

In a presentation made to the 2006 Annual Meeting of the Canadian Society for Agronomy, Zheljazkov et al showed that while application of linuron to dill caused about 30% plant mortality, this degree of damage had no impact on yields. This illustrates the yield plasticity of this crop. They identified a number of herbicides that were less toxic to the crop than linuron – these products should be considered for use in future work with dill.

By early July, the dill crop had out-grown the effects of the herbicide – except that the sprayed plants were considerably shorter than plants in areas missed by the spray. This is illustrated in Figure 2006-1. The crop was irrigated once in late July.

By the 3rd week in August, the flowers and some of the upper leaves on some of the lines had begun to yellow. Initially this was attributed to these lines being early maturing – but it turned out that the yellowing was largely due to disease. The disease started in a couple of localized points, but spread throughout the plot with a one week period. Although the causal organism was not isolated, the symptoms are typical of blossom blight. This disease complex has become fairly widespread in commercial plantings of coriander and caraway in Saskatchewan – but dill is usually not as severely affected. Distribution of the disease was not uniform across the plots – but whether this reflects differences in disease sensitivity of the various DH lines – or was due to differences in distribution on the initial disease inoculum could not be determined. Because of the rapid spread of the disease, no chemical control was attempted. Trials conducted by the University of Saskatchewan with blossom blight of coriander suggest that this disease cannot be controlled utilizing commercially available foliar-applied fungicides once it has become established within the canopy. Fungicidal seed treatments followed by a preventative spray program may be more effective – and this practice was followed in subsequent trials.

Each line was straight cut combined as it matured using a WinterSteiger Plot Master combine. Due to the indeterminate nature of dill, some pre-harvest shattering loss of the first seed set is inevitable. In 2006, the onset of wet, cool weather in mid-September resulted in a substantial delay in the harvest of many of the later maturing lines (September 26). Unfortunately, once dill is ready to the crop is ready to harvest, anything delay can result in substantial losses to shattering. Although swathing may reduce problems with shattering, most commercial growers try to avoid swathing dill because of its rank growth habit. It may be useful to try to chemically desiccate the later maturing lines.

The seed was air dried after harvest and then cleaned using a dockage tester.

Table 2006-1. Agronomic Observations on Double Haploid and Parental lines of Dill (2006).

Yields per unit area were below the average of 1000 kg/ha seen in most commercial dill fields – this likely reflects the fact that the rows were very widely spaced to facilitate roguing and crop evaluation. Of more importance is the fact that several of the DH lines had yields that were substantially greater than the industry standards (Mammoth and Giant). It is noteworthy that the yields exceeded that of Mammoth – as it was Mammoth that was used to generate the DH lines included in this test.

Oil content and quality were evaluated utilizing standard methods (Arganosa et al 1999). Briefly, a 5 g sample of cleaned seed was steam distilled to extract the essential oils and then the composition of the oil is determined by HPLC.

The results of the oil analysis and the corresponding oil yields are presented in Table 2006-2. Overall, the oil content was slightly higher than normal – this likely reflects the warm growing conditions through July and August of 2006. Although there was a significant degree of variation in the essential oil content of the various lines tested, the composition of the oil (limonene : carvone ratio) was quite similar (Table 2006-2).

Table 2006-2. Oil content and composition for double haploid and parental lines of dill (2006)

The essential oil content of DH 12 was superior to all other DH lines, as well as the industry standard lines.

Notable performance, yields and quality characteristics of the lines were;

a. Line DH-12 had performed very well in previous trials – but in this trial it appeared to experience a greater degree of damage by the herbicide than the other lines. This may reflect the fact that DH-12 emerges very quickly – resulting in this line being at a more advanced stage of crop development at the time of the herbicide was applied. This may have rendered it more susceptible to herbicide damage. DH-12 is also very uniform in height, with a short stature and early maturity (Table 2006-1). Figure 2006-3 shows that on July 20, DH-12 was in full bloom – whereas most of the other lines were just starting to bud out at that time. In the 2006 trial this meant that this line was flowering at a time when weather conditions appeared to be most suitable for development of the blossom blight. As a consequence many of the seeds formed by this line were shrunken and sterile. This is reflected in the exceptionally high grade out and low 1000 seed weight for DH-12. Yields for this line were consequently lower than normal. However, DH-12 is very early maturing. This is reflected in the exceptionally high oil content of seed of this line.

b. Line DH-47 is another short stature line that had performed well in previous trials. Although this line showed excellent seedling vigor, it was quite severely affected by blight, resulting in low seed yields and a high % seed grade out due to shrunken (blight affected) seeds (Table 2006-1). The poor quality of the seed harvested for this line was also reflected by its low oil content (Table 2006-2).

c. CDC Giant – experienced some of the same herbicide toxicity problems as D12. This line also appeared to be particularly sensitive to the blossom blight. The essential oil content of CDC Giant was quite high, reflecting the strong selection pressure for this trait in dill breeding programs.

d. Lines DH-7 and DH-53 – in this trial, as in previous trials, these lines were very vigorous, producing plants that were taller than the parental line (cv Mammoth) but much more uniform in height. Both of these lines were relatively late maturing. Harvest of these lines was delayed due to bad weather from mid-September onwards. This resulted in extensive seed losses to shattering. Despite this loss, these lines were the highest yielding in the trial. In part, this could be attributed to the fact that these lines were relatively unaffected by the blight. Whether this reflects inherent disease resistance or simply reflects differences in growing conditions at the key stages of development of both the crop and the disease infection could not be determined. Although these lines were both quite late, the essential content of the seed was acceptable – with DH 53 having a seed oil content superior to CDC Giant. Overall oil yield/unit area of Lines DH-7 and DH-53 were respectively 58% and 116% greater than the industry standard Mammoth – this reflects both the high yields and relatively high seed oil content of these lines.

e. Line DH-1 – in previous trials this line had been extremely late but had produced very large plants with very large seeds. Seed viability was very poor – likely reflecting problems with getting mature seed. The poor seed viability led to a poor stand and seed yields were consequently low. The seed oil content was also quite low.

Evaluation of New DH Dill Lines
Over the winter of 2005/2006 the program generated seed of the DH lines that had produced the best results in the 2004 and 2005 field trials (DH 12, 47 and 53). This step was considered prudent given the poor viability of the seed generated in the 2004 and 2005 field trials. This seed was used to establish a nursery for seed generation of these DH lines. Using transplants to establish the nursery insured that high quality mature seed could be harvested in a timely manner. It also allowed for evaluation of the germination %, field performance and seed quality of the selected lines against the parental lines. Two new DH lines were also included in this trial.

Results
The germination % of the previously selected DH lines was not as good as the parental types, but once in the field these DH lines again out-performed their corresponding parental lines, producing uniform stands of vigorous plants. DH 12 and 47 were far earlier to mature than the parental lines, while DH 53 produced a very vigorous, uniform stand of late maturing but high yielding plants. It is interesting to note that there was very little disease in this trial – while these same lines (DH 12 and DH 47) had been severely affected by disease in the adjacent direct seeded trial. This suggests that disease losses in this crop are strongly influenced by interactions between disease, crop development stage and environmental conditions. The two new DH lines included in this trial failed to germinate.

2007 – DH Dill Trials

The 2007 trial utilized the most promising DH lines selected from previous trials. Lines were included in these trials if they had demonstrated promise in any of the following areas;

a) plant vigor, uniformity of height, uniformity of maturity
b) seed yield
c) oil content of the seed
d) chemical composition of the seed oil.

Key objectives were to identify lines that retained the vigor of the parental lines, but were shorter and more uniform in stature. Improved uniformity of seed set was a highly desirable characteristic. The presently available commercial dill lines have a very indeterminate growth habit resulting in uneven seed maturity. With highly indeterminate seed set it is inevitable that a portion of the seed is lost to shattering before the crop is ready to harvest, while a portion of the seed is also still immature at harvest. Immature dill seed does not have the flavor profile required by the industry.

Unless otherwise specified, the procedures utilized in the 2007 DH dill trials corresponded to the procedures previously described for the 2006 trials. The trial was again conducted at the University of Saskatchewan Horticulture Research Facility in Saskatoon. The test site had been in pumpkins in 2006 which resulted in greater weed pressure than in previous years. The field was prepared as previously described. The trial was seeded on May 23. This is a later planting date than most commercial growers would use – but it reflects the fact that the clay soil at the test site is very slow to dry and warm in the spring. The crop was again seeded with a push-type small plot seeder. Each line was planted out in 3 rows (0.5 m apart) with each row being 8 m long. Each plot was separated by a 1 m buffer. The trial was laid out in a randomized complete block design with 4 replicates.

The seed used in the 2007 trials was generated in the 2006 trial. Although the germination % of the seed was relatively good, we still opted to seed relatively heavily because of the heavy nature of the soil at the test site. This resulted in a dense stand for most lines except DH 45 and 47 (Table 2007-1). By July, the crop canopy for the various DH dill lines appeared quite uniform – despite the difference in initial plant stand. This reflects dill’s indeterminent growth habit.

Warm weather coupled with timely rainfall in early June resulted in rapid and uniform crop emergence but also produced near-ideal conditions for the germination of weeds – particularly red root pigweed and common groundsel. Weed pressure was particularly heavy in the 4th block of this trial. In the 2006 trial the herbicide linuron (Lorox) applied at 200 g a.i./a once the crop reached the 3 true leaf stage had caused significant crop damage (yellowing and some seedling death). In the 2007 trial we again opted to use Linuron, but at a lower rate (150 g a i./a). The lower rate of linuron reduced but did not eliminate herbicide damage to the crop (see Table 2007-1) and also provided less effective weed control – especially in the heavy weed pressure in the 4th block. The 4th block had to be hand weeded on several occasions in 2007.

In the 2006 trial, the DH lines as well as the industry standard and parental lines had appeared uniformly susceptible to the relatively high rate of linuron. In the 2007 trial when the lower linuron rate was used there appeared to be some variability in relative sensitivity – with several of the DH lines showing more damage than the established lines (Table 2007-1). Whether these differences reflects actual differences in herbicide sensitivity or are simply related to differences in crop developmental stage at the time of exposure to the linuron could not be determined. The fact that CDC Mammoth seemed particularly resistant to linuron in 2007, yet in the 2006 trial it was amongst the most sensitive lines, suggests that the effect was mediated by environmental conditions or developmental stages.

By early July, the dill crop had out-grown the effects of the herbicide, but development and vigor of the 4th block was delayed by weed pressure. The crop was irrigated once in late June and again two weeks later. Although the soil was quite dry from mid-July onwards we opted to not irrigate because of the problems with blossom blight observed in the 2006 trial.

Because of the problems with blossom blight observed in 2006, we opted to implement a preventative spray program in 2007. The crop was sprayed with azoxystrobin (Quadris) or chlorthalonil (Bravo) every 10 days from mid-July as the crop began to come into flower through until mid-August by which time the crops had begun to mature. There were no obvious signs of blossom blight in 2007 – whether this indicates the spray program was effective could not be determined as there were no untreated areas for comparison. Although there was no disease apparent, some lines again flowered well yet set relatively few seeds and many of those seeds were shrunken and of low quality. Whether this problem reflects a disease or is a varietal response to adverse growing conditions at flowering could not be determined.

Seed loss due to shattering, both prior to harvest and during combining, is a major issue during the harvest of dill. In previous years the crop was straight cut combined once it had matured – but this resulted in significant shattering loss, particularly in early maturing lines or in situations where die-down was delayed by cold, wet weather. In 2007 we opted to chemically desiccate the crop with diquat (Reglone at 1.0 l/a) as soon as it began to mature, followed by hand swathing 5 days later to minimize shattering due to wind action. One week later the swaths were picked up and combined using a standard WinterSteiger small plot combine. Although this harvest procedure did not completely eliminate shattering loss, the losses were far lower than in previous years. This is reflected by the fact that yields in 2007 were almost 10X higher than in 2006 (Table 2007-1). Although the altered harvesting procedures may have made a significant contribution towards the higher yields observed in 2007, near ideal growing conditions through September followed by very favorable harvest conditions were also helpful.

Yields per unit area in 2007 were all well above the average of 1000 kg/ha seen in most commercial dill fields – this likely reflects the near ideal conditions in 2007, coupled with the extra care taken to reduce seed loss due to shattering.

In the 2006 trial several of the DH lines had yields that were substantially greater than the industry standards (CDC Giant) or the parental line (Mammoth). In 2007, none of the DH lines outperformed the standards. This may reflect the fact that growing conditions through the fall of 2007 were near ideal – this would have been advantageous for the relatively late maturing lines like CDC Giant and Mammoth. In years with an earlier or less favorable fall the earlier maturing DH lines would likely produce better yields and as the seeds would be more mature, the oil content would also expected to be higher.

The seed was air dried after harvest and then cleaned using a dockage tester. Quality analyses of the essential oils extracted from the seed were conducted as previously described.

Table 2007-1. Agronomic observations on double haploid and industry standard lines of dill in 2007).

Notable performance and yields characteristics of the DH dill lines tested in 2007 were;
a. Line DH-12 produced yields that were within 10% of the industry standard lines, yet it was ready for harvest more than a month earlier (August 23 versus October 6) than the standard lines. DH 12 again produced a vigorous stand of fast growing uniform plants. While it had showed significant herbicide damage in the 2006 trial, there was relatively little damage in 2007. This suggests the importance of developmental stage in determining sensitivity to linuron. DH 12 is far shorter in stature than the standard lines. A shorter plant stature may be desirable in rank crops like dill as it reduces the mass of material to be combined. In the 2006 trial, line DH 12 experienced severe yield loss to what was thought to be blossom blight. In the 2007 trial no blossom blight was observed, yet there was again a fairly high incidence of aborted or shrunken seed in DH 12. Flowering of DH 12 is quite synchronous and occurs when the plants are relatively small. It is possible that the apparent issues with seed set simply reflect the plants’ inability to fill the huge number of seeds that set at one time. Selecting for synchronous flowering at an early growth stage may be a yield limiting strategy – however it tends to guarantee at least some yield. Seed of DH 12 was quite large, again suggesting a high degree of maturity.
b. Line DH-47 is another short stature line that had performed well in early trials. However the seed of DH 47 generated in 2006 had been of poor quality and that resulted in a poor stand and limited yields in the 2007 trial.
c. CDC Giant – had problems with herbicide toxicity and blight in 2006, but in 2007 it produced the highest yields of the lines tested.
d. Lines DH-7 and DH-53 – in 2007 these lines again produced vigorous stands of uniformly tall plants. Both of these lines were relatively late maturing and their yields were substantially lower than industry standards which are also late maturing. DH-7 produced the largest seeds of any line tested in 2007.

Seed Quality in 2007 (see Table 2007-1)
Overall the essential oil content of the seed harvested in 2007 was slightly higher than in 2006, reflecting the greater degree of crop maturity achieved in 2007. The seed harvested in 2007 had a higher limonene content than seen in 2006 – the reason for this quality shift is not clear. Within the lines tested in 2007, there was less variability in the oil composition (% limonene) than in the oil content. Oil extracted from CDC Giant appeared to have a lower limonene content than any of the other lines tested in 2007. Oil yields (kg/ha) were far higher in 2007 than in 2006. This reflects the higher seed yields in 2007 rather than any improvement in seed oil content.

While line DH-12 had the highest seed oil content in the 2006 trial, its seed oil content was amongst the lowest of the lines tested in 2007. This emphasizes the importance of multi-year trials when evaluating any germplasm. DH-53, which is quite late maturing, had the highest seed oil content in both the 2006 and 2007 trials.

2008 – DH Dill Trials
The 2008 trial focused on the DH lines tested in 2007 as no new lines tested in 2007 had shown sufficient promise to merit inclusion in this trail.

Unless otherwise specified, the procedures utilized in the 2008 DH dill trials corresponded to the procedures previously described for the 2007 trials. The trial was again conducted at the University of Saskatchewan Horticulture Research Facility in Saskatoon. The test site had been in lettuce and onions in 2007 – this resulted in relatively limited weed pressure – except for common groundsel which has become problematic throughout the test plots. The field was prepared for planting as previously described. The trial was seeded in the 3rd week of May. This is a later planting date than most commercial growers would use – but it reflects the fact that the clay soil at the test site is very slow to dry and warm in the spring. The crop was again seeded with a push-type small plot seeder. Each line was planted out in blocks of 4 rows, with each row being 6 m long. The between row spacing in the 2008 trial was reduced to 0.3 m as compared to 0.5 M in previous years. The tighter row spacing was used to produce a thicker crop canopy earlier in the season, thereby potentially reducing problems with weed competition. Each plot was separated by a 1 m buffer. The trial was laid out in a randomized complete block design with 4 replicates.

The seed used in the 2008 trials was generated in the 2007 trial. Although the germination % of the seed was relatively good, we still opted to seed relatively heavily because of the heavy nature of the soil at the test site. This resulted in a dense stand for most lines except DH 7 and 45. By July, the crop canopy for the various DH dill lines appeared quite uniform – despite the difference in initial plant stand. This reflects dill’s indeterminent growth habit.

Cool weather through late May and most of June resulted in extremely slow and uneven crop emergence. Some plants came up within 2 weeks of planting, but the majority took over a month to emerge. This delayed emergence response appeared to be consistent across lines and replicates. Although conditions were not favorable for emergence of the dill crop, they were suitable for the germination of weeds – particularly common groundsel. A healthy weed population coupled with slow crop emergence resulted in heavy weed pressure throughout the plot area. In the 2006 trial the herbicide linuron (Lorox) applied at 200 g a.i./a once the crop reached the 3 true leaf stage had caused significant crop damage. In the 2007 trial using Linuron at a lower rate (150 g a i./a) had reduced crop damage but also provided less effective weed control. Because of the heavy weed pressure in 2008 we opted to return to the higher rate of Linuron. This resulted in little crop damage and decent weed control, except for the common groundsel. The groundsel was so advanced by the time that the dill crop was finally ready to spray that the groundsel went to seed before the herbicide became effective. As these weed plants collapsed and died they released their seeds. These seeds were incorporated into the soil during the supplementary tillage operations required to clean up weed escapes. These weed seeds germinated following each rain event – creating flush after flush for the duration of the growing season. These weeds were controlled by hand tillage until the dill crop grew to the point where it shaded out the small stature groundsel plants.

Because of the problems with blossom blight observed in 2006, we again opted to implement a preventative spray program in 2008. The crop was sprayed with azoxystrobin (Quadris) or clorthalonil (Bravo) every 10 days from mid-July as the crop began to come into flower through until mid-August by which time the crops had begun to mature. There were no obvious signs of blossom blight in 2008 – whether this indicates the spray program was effective could not be determined as there were no untreated areas for comparison. Although there was no disease apparent, some lines again flowered well, yet set relatively few seeds and many of those seeds were shrunken and of low quality. Whether this problem reflects a disease or is a varietal response to adverse growing conditions at flowering could not be determined.

Seed loss due to shattering, both prior to harvest and during combining, is a major issue during the harvest of dill. In 2006 the crop was straight cut combined once it had matured – but this resulted in significant shattering loss, particularly in early maturing lines or in situations where die-down was delayed by cold, wet weather. In 2007 we had tried chemically desiccating the crop, followed by swathing 5 days later and then combining after another week of dry down. Although this harvest procedure did not completely eliminate shattering loss, the losses were far lower than in previous years. In 2008 we tried to further refine this process. We chemically desiccated the crop as before, but allowed it to dry standing and then straight combined the crop. The thought was that the standing crop would dry out more quickly and that elimination of the swathing step might also reduce total loss to shattering. Unfortunately conditions in the fall were less than ideal for this approach to crop management. All lines were slow to mature in 2008 and therefore the desiccation and dry down step occurred during relatively cool weather. This slowed the drying process, leaving the standing crop exposed to shattering loss for extended periods of time. This problem was most severe on the late maturing lines (DH 7 and 53) which stood for more than 3 weeks before they were dry enough to combine. Although it was not possible to quantify shattering losses, visual inspection of the field suggested that it was substantial.

Despite problems with weed competition and shattering losses, seed yields in 2008 were all well above the average of 1000 kg/ha seen in most commercial dill fields.

In the 2006 trial several of the DH lines had yields that were substantially greater than the industry standards (CDC Giant) or the parental line (Mammoth). In 2007, none of the DH lines outperformed the standards. In the 2008 trial, the early maturing DH lines (DH 12 and 35) had yields that were significantly higher than the standards. Yields of the later maturing lines (DH 7 and 53) were reduced due to shattering loss.

The seed was air dried after harvest and then cleaned using a dockage tester. Quality analyses of the essential oils extracted from the seed were conducted as previously described.

Table 2008-1. Agronomic observations on double haploid and industry standard lines of dill in 2008).

Notable performance and yield characteristics of the DH dill lines tested in 2008 were;

a. DH 35 has emerged as the line that most consistently delivers yields superior to the existing standard lines. It produces a vigorous stand of moderate stature plants that are ready to harvest about 1 week before the standard lines.

b. Line DH-12 produced yields in 2008 that exceeded the industry standard lines, and it was ready for harvest two weeks earlier (Sept 10 versus October 1) than the standard lines. In previous trials DH 12 had shown superior vigor early in the season. This was also apparent in 2008 – where a portion of the seeds germinated within 2 weeks of planting despite less than ideal field conditions. Although these plants got off to an early start, the majority of the DH seedlings took over a month to emerge. This resulted in a very uneven crop for DH 12 in 2008 – whereas in other years this line had been exceptionally uniform. Seed from the early emerging plants of DH 12 was lost to shattering before the rest of the plants in the DH 12 plots were ready to be desiccated.

c. Line DH-47 is another short stature line that had performed well in 2008. This line had produced a poor stand in the 2007 trial but produced an acceptable stand in 2008.

d. As usual, the DH-7 and DH-53 lines produced vigorous stands of uniformly tall plants. Both of these lines were relatively late maturing and their yields were reduced due to shattering loss in 2008.

Seed Quality in 2008 (see Table 2008-1)
Overall the essential oil content of the seed harvested in 2008 was slightly lower than in 2007, but was comparable to 2006 – this reflects the relatively greater degree of crop maturity achieved during the exceptionally long and warm 2007 growing season. The seed harvested in 2008 had a higher limonene content than seen in 2006 or 2007 – the reason for this shift is not clear. Within the lines tested in 2008, there was again less variability in the oil composition (% limonene) than in the oil content.

Oil extracted from CDC Giant appeared to have a lower limonene content than any of the other lines tested in both 2007 and 2008. Oil yields (kg/ha) in 2008 were lower than in 2007 but higher than in 2006. Year to year differences in seed yield have a greater impact on oil yields/unit area than differences in seed oil content. . DH-12 had the highest seed oil content in the 2006 and 2008 trials, but its seed oil content was amongst the lowest of the lines tested in 2007. DH-53, which is quite late maturing, had the highest seed oil content in both the 2006 and 2007 trials, but in the 2008 trial its seed oil content was near the mean. These year to year differences in performance emphasize the importance of multi-year trials when evaluating any germplasm.

Conclusions for DH Dill and Caraway Trials
In replicated field trials of DH dill conducted from 2006 through 2008 this project identified several lines that differed greatly from the parental lines in important agronomic characteristics such as speed of crop emergence, crop stature and time to maturity. These traits may be of value in a breeding program, especially as they are present in a homozygous form in the DH lines. We also identified several DH lines that were superior in seed and seed oil yield relative to the parental line and the most widely grown commercial dill variety. These lines may have immediate market value.

Opportunities for commercial release of the most promising DH lines identified in this project are being pursued by the Crop Development Center of the University of Saskatchewan working in collaboration with PBI/NRC and the Saskatchewan Herb and Spice Growers Association. The DH dill lines with traits of potential value in crop improvement programs will be preserved and made more widely available through the Plant Gene Resource Center of Agriculture Canada.

2006 – DH Caraway Trials

- pdf

A total of 25 DH lines of caraway were available for the 2006 trials – along with 3 parental lines. Because of limited amounts of seed available for the DH lines, all lines were seeded in the Dept of Plant Sciences greenhouses and then transplanted out in early June. The seedlings were raised in 144 cell transplant flats filled with Sunshine Mix #4 media. The flats were maintained under near optimum conditions for germination – a 24/18 C temperature regime with a 16h photoperiod. When the seedlings were 4 weeks old they were transplanted into the previously described field plot area (15 cm apart with 0.5 m between rows). The growth habit of the plants was observed. The trial was hand harvest in early October and then threshed using a stationary combine. The seed was cleaned using a dockage tester. Because of the uneven numbers of plants no yield comparisons were possible.

General observations
The parental material showed relatively poor seed viability even under the near-ideal conditions of the greenhouse (Table 2006-3). Once transplanted into the field, the parental lines performed relatively well – they appeared vigorous and were quite uniform in stature and maturity. No problems with disease were observed in the parental or the DH lines.

As has been seen in other crops, many of the DH caraway lines had serious agronomic drawbacks – ie; many had low germinate % even under ideal greenhouse conditions (Table 2006-3), or they had very poor vigor or an exceptionally short or distorted growth habit. Although some of these poorly adapted lines did produce seed, they will likely be omitted from future trials. Some of the DH lines (ie; DH-10 and 29) showed much better germination % than the corresponding parental lines. None of the DH caraway lines were obviously superior to the parental lines in terms of vigor, uniformity of plant configuration or yield.

Seed quality assessments were conducted on 5 g samples of clean seed, utilizing the procedures outlined in the dill section. The essential oil content of line DH-21 was significantly higher than the parental line, while for lines DH-14 and 29 the essential oil content of the seed was far lower than the parent. The oil composition of these lines was also quite different from the parent. The combination of low oil content and different oil composition may reflect the fact that the seed of these lines was not fully mature at harvest.

Sufficient seed was generated in this trial to allow a replicated yield trial to be conducted in 2007.

2007 – DH Caraway Trials
The 2006 field trials with DH annual caraway had produced very limited seed yields, and the seed was of poor quality. This strongly suggests that the growing season in Saskatchewan is not long enough to consistently mature the presently available lines of annual caraway. It is noteworthy that the 2006 crop failed to mature despite the fact that the crop received a head-start by transplanting coupled with the fact that 2006 was an exceptionally warm year.

The 2007 field trial with DH caraway was conducted adjacent to the DH dill trial, at the previous described University of Saskatchewan research site. All variables and procedures were identical to those described in the dill trial except;

- the trial was seeded using the DH caraway lines that had produced appreciable amounts of seed in 2006. This process strongly selected for fast maturing lines.
- due to limited quantities of available seed. the trial was only planted out in three replicates and each replicate consisted of only two rows rather than the three rows used in the dill trial.
- as annual caraway is extremely late maturing, the crop was left in the field until mid-October. It was then direct combined. The delayed harvest may have resulted in some seed loss to shattering. Delaying the harvest until this point in the season would be extremely risky for a commercial grower.

Results
The caraway was slower to emerge than the dill and the stand quality of the various lines was highly variable (Table 2007-2). Stand counts for the parental lines were not appreciably different than for DH 10 and 29. The caraway appeared to suffer fewer ill effects from the linuron spray than the dill. This may reflect the smaller size of the caraway plants at the time of spraying

In mid-June it became apparent that there was a severe problem with aster yellows in the entire research plot area – including the caraway plots. We had seen increasing levels of this disease in 2006, but did not suspect that the problem would persist into 2007. Nonetheless, the caraway crop was heavily infected and the effect on the plants was much more severe than in dill or coriander – and much more like the devastating effect seen in carrots. The plants became yellow and produced multitudes of small distorted leaves and stems. None of the infected plants survived to produce seed and as a consequence, seed yields in this trial were compromised. There did not appear to be any significant difference in the incidence or impact of aster yellows for the various DH lines or the parental lines of caraway.

Given the problems with stand establishment, and aster yellows, seed yields were actually unexpectedly high – exceeding by a substantial margin the yields typically seen under commercial production. It is risky to extrapolate small plot yields to field scale operations – but the results are promising. Of particular interest is the fact that line DH 10 out-yielded the line it was derived from (NN-2) by a factor of 5 fold. In both the 2006 and 2007 trials, DH 10 had produced a better stand than its parent. In the 2007 trial, this stand advantage was further enhanced by the fact that DH 10 appeared to be less susceptible to aster yellows than NN-2. Of note is the fact that the seed of DH 10 is substantially smaller than the seed of any of the other DH lines or the parental lines. This may be undesirable for sales direct to the consumer where large seed size is equated with quality – however it would be of little importance to the processing sector.

Quality analyses of the 2007 caraway crop were completed as planned. The seed oil content was much higher in 2007 than in 2006 – this likely reflects the fact that the crop was allowed to mature late into the fall of 2007. The seed oil content of the DH lines was comparable to the parental lines – although the variability between lines was greater than that seen in the dill trial. Again the oil composition (% limonene) was more stable across the lines tested than the actual oil content. The highest yielding DH caraway line (DH 10) had a moderate seed oil content – but for total oil produced/ha it would clearly have exceeded all other lines, including its parent.

2008 – DH Caraway Trials

The 2008 field trial with DH caraway was conducted adjacent to the DH dill trial, at the previous described University of Saskatchewan research site. All variables and procedures were identical to those described in the dill trial except;

- the trial was seeded using the DH caraway lines that had produced appreciable amounts of seed in 2006 and 2007. This process strongly selected for fast maturing lines.
- as annual caraway is extremely late maturing, the 2008 caraway crop was left until mid-September before it was desiccated and harvest was delayed until mid-October. This process may have resulted in some seed loss to shattering. Delaying the harvest until this point in the season would be extremely risky for a commercial grower.

Results

The caraway was even slower to emerge than the dill in 2008 and as noted with the dill this led to problems with weed competition. Stand counts for the parental lines were not appreciably different than for DH 10 and 29 (data not shown). As noted in previous years, the caraway appeared to be more tolerant of the linuron spray than dill. This may reflect the smaller size of the caraway plants at the time of spraying.

As with the 2008 dill crop, the linuron treatment of the 2008 caraway plot came too late to provide adequate control of the dominant weed in the plot area – common groundsel. While the dill crop outgrew the relatively short stature groundsel plants by mid-July, the shorter caraway plants had to be hand weed for the duration of the 2008 crop season.

Aster yellows was less of a problem in the 2008 caraway crop than in 2007, with only about 10 % of the plants showing signs of infection. As the infected plants are too small and chlorotic to produce a significant number of seeds, yields may have been reduced due to this problem.

Given the problems with stand establishment and weed control, seed yields in 2008 were actually unexpectedly high – exceeding by a substantial margin the yields typically seen under commercial production. Line DH 10 again out-yielded the line it was derived from (NN-2), although by a far smaller amount than was observed in 2007.

Quality analyses of the 2008 caraway crop were completed as planned. The seed oil content was higher in 2008, than in either 2006 or 2007 – this likely reflects the fact that the crop was allowed to mature late into the fall of 2008. The seed oil content of the DH lines was higher than the parental line (NN-2). Again the oil composition (% limonene) was more stable across the lines tested than the actual oil content. The highest yielding DH caraway line (DH 10) in the 2008 trial also had the highest seed oil content. Oil yield (kg/ha) for the best DH caraway line (DH10) was 50% higher than for the parental line (NN-2).

Conclusions for DH Caraway Trials

In replicated field trials of DH caraway conducted from 2006 through 2008 this project identified one line of DH annual caraway (DH-10) that was clearly superior to presently grown varieties for both seed yields and quality. However this line still required a longer growing season than is available in Saskatchewan – and therefore the market for this improved DH line of annual caraway lies elsewhere. Opportunities for commercial release of the most promising DH lines identified in this project are being pursued by the Crop Development Center of the University of Saskatchewan working in collaboration with PBI/NRC and the Saskatchewan Herb and Spice Growers Association. The DH caraway lines with traits of potential value in crop improvement programs will be preserved and made more widely available through the Plant Gene Resource Center of Agriculture Canada.

Field evaluation of doubled haploid plants in the Apiaceae: dill (Anethum graveolens L.), caraway (Carum carvi L.), and fennel (Foeniculum vulgare Mill.)

A. M. R. Ferrie • T. D. Bethune
G. C. Arganosa • D. Waterer
-pdf

Abstract

The Apiaceae family includes vegetables, as well as herb and spice crops. Compared to major crops, there have been few breeding or genetic improvement programs for any of the Apiaceae, especially the herb and spice species. Haploidy technology can be used to develop uniform, true-breeding lines, as well as to accelerate breeding programs.
Field trials of dill (Anethum graveolens L.), caraway (Carum carvi L.), and fennel (Foeniculum vulgare Mill.) doubled haploid (DH) lines were conducted over 2–5 cropping seasons. Several of the DH dill lines had desirable agronomic characteristics such as short uniform stature along with early maturity that would be useful for crop improvement. Seed yields and the essential oil content of the seed harvested from the best performing DH dill lines were either equal to or higher than the parental line.

A DH annual caraway line was identified that produced higher seed yields than the industry standard. The main constituents of the essential oil for the DH lines of both dill and caraway were similar to the parental lines. Fennel DH lines exhibited differences in height but were too late in maturity for seed production under prairie conditions. The results indicate that not only were we able to generate DH lines that could be used in a crop improvement program, but we developed DH lines that could be used directly as cultivars as these lines performed better than the industry standard (parental line).

Introduction
The Apiaceae include a number of economically important vegetable, herb, and spice crops used as food, flavourings, perfumes, cosmetics, and medicinals for humans and animals (French 1971). They are grown worldwide, but are most common in the temperate regions of the northern hemisphere. There is an economic need to develop new cultivars of Apiaceae species better suited to Canadian prairie conditions. Earlier maturity coupled with increased yields of fruit (commonly known as seed), higher seed oil content, and better oil quality are the main objectives of improvement programs for Apiaceae.


Dill (Anethum graveolens L.) is an annual or biennial herb used primarily as a condiment. Dill seed and leaves are used as flavouring in sauces, vinegars, pastries, and soups. Dill has medicinal value as a diuretic, stimulant, and a carminative (Peirce 1999). The main constituents of the oil extracted from dill seed are D-carvone (35–60%) and D-limonene (35–50%) (Weiss 2002). Caraway (Carum carvi L.) is an annual or biennial herb native to Europe or western Asia. Caraway seed is used whole as a spice or crushed to produce caraway oil. The seeds have a licorice flavour and are used in breads, soups, spreads, salad dressings, and liqueurs. The leaves can be used in cooking, as can the roots. Caraway seed and seed oil have medicinal applications for disorders such as rheumatism, eye infections, toothaches, and nausea (Peirce 1999). The main constituents of the seed oil of caraway are similar to dill with D-carvone (45–60%) and D-limonene (35–55%) (Weiss 2002). Carvone has been used as a flavour additive in foods, a sprouting inhibitor for potatoes (Beveridge et al. 1981), a growth inhibitor for fungi (Smid et al. 1995) and bacteria (Helander et al. 1998; Oosterhaven et al. 1996) and as an insect repellent (Salom et al. 1996; Lee et al. 1997). The oil can also be used as a fragrance component in cosmetics (e.g., soaps, creams, lotions, and perfumes). Fennel is an erect, umbelliferous herb of the same family (French 1971). There are three subspecies of Foeniculum vulgare ssp. capillaceum: azoricum, also known as bulb fennel, Italian fennel, or Florence fennel; dulce, also known as sweet fennel, French fennel, or oilseed fennel; and vulgare (bitter fennel), which is perennial, and has an essential oil content higher than that of dulce. Bulb fennel is used as a vegetable, whereas sweet fennel is used as a condiment. Fennel seed is used in the food industry to flavour meats, vegetables, fish, soups, salad dressings, stews, breads, pastries, teas, and alcoholic beverages. Trans-anethole is the dominant component in the essential oil extracted from fennel seeds (Weiss 2002). This oil is used in condiments, soaps, creams, and perfumes. The medicinal or nutraceutical applications of fennel include use as an antispasmodic, carminative, diuretic, expectorant, laxative, and stomachic (Peirce 1999). Doubled haploid (DH) technologies, to produce embryos/plants from haploid immature pollen grains (microspores), are being used around the world to develop new cultivars and uniform lines in many plant species (Thomas et al. 2003). The main advantage to plant breeders of using doubled haploid breeding lines is the reduction in time required to achieve homozygosity. True-breeding lines can be generated in one generation rather than several years of backcrossing or selfing. This is also beneficial for producers and consumers as high yielding cultivars with improved agronomic and quality traits can be developed more rapidly. The objective of this project was to compare the agronomic performance of DH dill, caraway, and fennel lines derived via microspore culture against their corresponding parental lines under field conditions. Materials and methods The DH lines were derived from locally adapted cultivars using the doubled haploidy methodology outlined in Ferrie et al. (2010). As seed supply of the DH lines was limited, lines were initially grown in small plot non-replicated trials. The seed obtained from these trials was used in subsequent larger scale replicated trials. The agronomic characteristics of the DH lines were evaluated in field trials conducted at the University of Saskatchewan Horticulture Research Facility in Saskatoon, Saskatchewan, Canada. Unless otherwise specified, the crops were managed using standard commercial production practices. Soil fertility in the plot area was adjusted prior to planting to meet industry recommendations. Weeds were controlled using tillage and herbicides. The plots were only irrigated under extreme drought conditions to improve crop emergence or as required to activate soil-applied herbicides. No insect or disease control measures were employed.

Dill
A total of 41 DH lines of dill were evaluated from 2003 to 2008. Any DH line with poor seed viability (\20% germination) was eliminated from further field testing. In 2003, viable seed of 12 DH lines was available to conduct a field trial. Due to the limited seed supply and unknown vigor of the DH lines, the seedlings were initially grown for 6 weeks under greenhouse conditions before being moved to the field. The seedlings were planted out in the field in late May in rows spaced 0.5 m apart, with 15 cm between plants within the row. Seedlings of the parental line (‘Mammoth’) were included in this trial for comparison purposes. In 2004, the DH dill lines that had produced the most complete stand and the highest seed yields in the 2003 field trial were evaluated in a direct-seeded field trial that again included the parental line. This trial was initiated using seed harvested from the field trial conducted in the previous year. The trial was planted in mid-May using a small plot seeder. Each plot consisted of four 8 m long rows of each DH line. The trial was laid out in a randomized complete block design with three replicates. Seed was harvested as each line matured using a small plot combine. The 2005 trial was lost to flooding. Field evaluations were repeated in 2006, 2007, and 2008 using the same procedures, except in 2007 when each line was chemically desiccated with diquat (Reglone) applied at 2.2 l/ha as soon as the line began to mature. The lines were hand swathed 5 days later, allowed to dry in the field for 1 week, and then combined as previously described. This modification to the harvest method was implemented in an effort to reduce the shattering losses that occur as the dill crop begins to ripen. In each year of testing, the DH lines and the parental line were compared for time of bloom, plant height at maturity, seed yields, seed oil content, and the relative amounts of the main constituents in the oil.

Caraway
A total of 25 DH lines of annual-type caraway and three parental lines (NN-1, NN-2, and Moran) were available for field testing, beginning in 2006. As with dill, the first year of testing of caraway involved a limited number of plants established as transplants in a non-replicated trial. The 4-week old transplants were set out in rows spaced 0.5 m apart with 15 cm between plants within each row. The trial was hand-harvested in early October and threshed using a stationary combine. Because of the highly variable numbers of plants for each line, no yield comparisons were possible in 2006. The three DH caraway lines that had produced significant quantities of seed in the 2006 trial were evaluated in a replicated direct-seeded field trial in 2007 and 2008. These trials were planted and maintained as described in the dill trial. The trials were chemically desiccated in early October at which time all of the lines were still growing vigorously. Due to the short stature of the caraway plants, they had to be hand harvested and then threshed using a stationary combine.

Fennel
The fennel trials were conducted in 2003 and 2004. Due to limited seed supplies, all lines were grown from transplants rather than by direct seeding. In late May, six-week old fennel seedlings were transplanted into the field. Plants for each line were spaced 30 cm apart within the row, with 50 cm between rows. Each line was grown in a single 10 m long row. In the 2003 trial, 14 DH oilseed fennel lines were evaluated, along with the parental line. In 2004, another 11 DH oilseed fennel lines were evaluated along with the parental line. Germination percentages and plant heights were recorded. In both years, the growing season was too short to allow development of seed in any of the fennel lines tested.

Chemical analyses: dill and caraway
Essential oils were extracted from the harvested seed of dill and caraway by hydro-distillation using a Clevenger-type apparatus. Five grams of clean seed was ground for 40 s in a Krups coffee grinder, transferred into a 1,000 ml round bottom flask, after which 500 ml distilled water was added. Distillation was allowed to proceed for 2 h after the first few drops of distillate were obtained. After cooling the apparatus for 20 min, the length of the essential oil column in the side-arm of the Clevenger apparatus was measured and the volume of essential oil calculated. The results are expressed in percent essential oil (v/w) based on the weight of the air-dried seed. The oil was stored at -18_C in the dark prior to subsequent chromatographic analysis. A Hewlett-Packard gas chromatograph (Series Model 5880A) equipped with a flame ionization detector and a capillary column containing 95% dimethyl-5% diphenyl polysiloxane (DB-5; 10 m 9 0.25 mm i.d.; film thickness: 0.25 lm; J&W Scientific, Folsom, CA) was used. Helium was the carrier gas with a column inlet pressure of 2.1 kg/cm2. The injector and detector temperatures were set at 275 and 300_C, respectively. The oven temperature was programmed as follows: [a] initial temperature: 75_C; [b] initial time: 5.0 min; [c] program rate: 8_C/min; [d] final temperature: 200_C; and [e] final time: 2 min. Ten microliters of essential oil were diluted to 500 ll with chromatographic grade hexane prior to gas chromatography. One microliter was injected at a split ratio of 1:50. The individual peak areas were quantified using a Hewlett-Packard 3396 Series II integrator with a chart speed of 1.0 cm/min and an attenuation of 2. The retention times of the compounds of interest, carvone and limonene, were compared to the retention times of a pure carvone and limonene standard runs under identical conditions. The results were expressed as a percentage of the total peak area.

Statistical analysis
All data from the replicated trials were analyzed using analysis of variance procedures. Where significant (P\0.05) F-test values were observed, means were compared using Fishers protected least significant difference test (P\0.05).

Results and discussion

Dill
The presently available commercial dill lines have an indeterminate growth habit resulting in uneven crop maturity and challenging harvest conditions. With highly indeterminate seed set, it is inevitable that a portion of the dill seed is lost to shattering before the crop is ready to harvest, while another portion of the seed is still immature at harvest. Immature dill seed does not have the oil content or flavor profile required by the industry. The development of an early, uniform-maturing dill line would be beneficial for producers.
For 23 of the 41 DH dill lines tested, seed viability was extremely low (\20%) and those lines were eliminated from further testing (data not shown). A few of the DH lines showed a moderate degree of reduction of seed viability (40–60% germination) relative to the parental line ([90% germination). The poor germination rates of these lines limited their yield potential and these lines were also removed from the trial. The germination rates of the remaining DH lines were comparable to the parental line in all years of testing. Differences in plant height at maturity were observed between the parental line and the dill DH lines.

Over 4 years (2003, 2004, 2006, 2007), five of the DH lines (DH-7, DH-12, DH-35, DH-45, DH-47) were shorter than the parental cultivar Mammoth by 12–32%. Two lines (DH-1, DH-53) were similar in height to the parental cultivar. Irrespective of the average height, the DH lines were more uniform in stature than the parental line (Fig. 1). The height of the umbels in the DH dill lines also tended to be more uniform than in the parental line, which should improve the efficiency of mechanical harvesting. The shorter stature DH lines could be useful in situations where lodging leads to harvest problems; however, a taller stature might indicate greater yield potential if the leafy portion of the crop were destined for processing as dill weed. The dill doubled haploid line DH-1 is a very vegetative plant with large seeds. Seed production by this line was low because of the late maturity (Table 1), and the vegetative growth of the plant. This line might prove useful for dill weed and/or dill weed oil production, however the yield and oil content of the leaves were not analyzed in this project.
This line was not evaluated in 2007 and 2008 because of the low seed yield. Our selection focus was on early maturing lines with high seed yield. Differences were observed among the DH dill lines and the parental cultivar for date of flowering. On July 20, 2006, DH-12 was in full bloom, whereas most of the other lines were just starting to flower (3% flowering with DH-53 to 47% flowering in DH-45; Table 1). The parental line was 78% flowering. There was year to year variation in time of flowering. On July 18, 2007, only two lines had started flowering, DH-12 and the parental line. In 2008, 60% of DH-12 was flowering, whereas other lines had not begun to flower (DH-7, DH-53), even though data was collected 8–10 days later than the previous years (Table 1). Early flowering and early maturity are beneficial for prairie climatic conditions, however it can also be detrimental as the earlier flowering lines appeared more prone to infection with blossom blight, as was the situation in 2006. As a consequence of this disease, many of the seeds were shrunken or non-viable. Relative seed yield of the DH dill lines, compared to the parental line, varied from year to year. In 2003, seed yields from the most productive DH lines (DH-7, DH-12, DH-35, and DH-45) were on average 61% higher than for the parental line (data not shown). In 2004, seed yields of DH-12 were far higher than the parental line (350% of Mammoth), with DH-35 also producing a high yield (139% of Mammoth) (data not shown). A frost in mid-August of 2004 severely damaged the later maturing DH lines, as well as the parental line. Both DH-12 and DH-35 are relatively early maturing, and they largely avoided the damaging effects of the frost. In 2006, several of the DH lines again had seed yields that were similar to or greater than the parental line; however, seed yields for DH-12 were poor in 2006 as this line was severely infested with blossom blight (Table 2). Seed yields in the 2007 trial were almost 10 fold higher than in previous years, indicating the magnitude of shattering loss that occurred when the dill was directly combined. None of the DH lines had seed yields that exceeded the parental line in 2007, although yields for DH-7, DH-12, and DH-35 were not significantly different from the parental line. The combination of more careful harvest management practices, coupled with near ideal conditions through the fall of 2007 was advantageous to the relatively late maturing parental line. In 2008, DH-35 was the highest yielding line and was ready to harvest 1 week before the parental line. DH-12 also produced a yield which exceeded Mammoth in 2008 and it was ready to harvest 2 weeks earlier than the parental line. In years with an earlier or less favorable fall, the earlier maturing DH lines would likely produce better yields and as the seeds would be more mature, higher oil content would be expected.
Over the 3 years of replicated trials, DH-12 and DH-35 had an average seed yield greater than the parental line. The seed yields obtained in these experiments were substantially greater than those obtained in previous studies. Bailer et al. (2001) reported yields of 400–600 kg/ha in 1 year and less than 200 kg/ha in another year. These low yields were due to seed shattering, a common problem with dill cultivars. A uniformly maturing line would alleviate some of the seed shattering associated with this species. Essential oil content of dill seed typically ranges from 1.2 to 7.7%, depending on the variety, growing conditions, and method of oil extraction (Weiss 2002). In Canada, seed oil content for dill typically ranges from 2 to 4%. Seed oil content of the DH dill lines evaluated in this project ranged from 2.2 to 3.7% over 3 years of replicated testing (2006–2008), with limited year to year variability in the oil content (Table 2). Over the 3 years, DH-12 consistently had similar or greater percent essential oil content compared to the parental line Mammoth. For most lines, the seed oil contained about equal amounts of carvone and limonene (Table 2) and this ratio was also consistent across production years. The North American dill seed oil industry is based on carvone content and stipulates that dill oil must contain at least 30% carvone; all DH dill lines tested in this trial exceeded this standard.

Caraway

In western Canada, caraway is typically grown as a biennial crop, as the growing season is too short for most presently available annual-type lines. The current annual types mature in about 120–130 days. Growing caraway as a biennial effectively ties up the field for two seasons, at significant cost to the grower. If annual types suited to short seasons were to become available they would represent a means to double the growers’ land use efficiency. A total of 25 DH lines and three parental lines of annualtype caraway were evaluated in field trials conducted in 2006. Seed viability was poor for most DH lines, with only three of the DH lines (DH-10, DH-21, and DH-29) showing more than 25% emergence of the planted seed.

However, it should be noted that the parental lines also had poor emergence percentages resulting in poor stand (Table 3). In 2007, the only DH caraway seed available for larger scale field trials was from the three DH lines that had performed relatively well in 2006. In both the 2006 and 2007 trials, DH-10 produced a better stand than its parent (NN-2; Table 3). In mid-June 2007, a problem with aster yellows disease developed in the DH caraway plots. The extent of the infection and the impact on the plants was much more severe than in adjacent plots of DH dill. None of the infected caraway plants survived to produce seed and consequently, seed yields were potentially compromised. There did not appear to be any significant difference in the incidence or impact of aster yellows on the various DH lines or the parental lines of caraway. Given the limited stand and the problems with aster yellows, seed yields were still unexpectedly high in the 2007 trial (Table 3). DH-10 produced 500% more seed than its parental line (NN-2) in 2007. In 2008, DH-10 again outperformed the parental line.
Seed of DH-10 is substantially smaller than the seed of any of the other DH lines or the parental lines (data not shown). This may be undesirable for direct sales to the consumer where large seed size is equated with quality, however seed size would be of little importance to the processing sector. Essential oil content of annual caraway seed typically ranges from 1.5 to 5.0%, depending on the variety, growing conditions, and method of oil extraction (Weiss 2002). The seed oil content of the caraway DH lines was much higher in 2007 and 2008 than in 2006 (Table 4); this likely reflects the fact that the crop was allowed to mature late into the fall of 2007 and 2008. The seed oil content of the parental line Moran was higher than any of the DH lines in 2007, however, this was not the case in 2006 or 2008. Over 90% of caraway seed oil consists of two monoterpenes, carvone and limonene. For the caraway DH lines, the ratio of carvone to limonene varied more between years than between lines in a given year. In Saskatchewan, carvone levels are around 46–50% in annual caraway and 54–57% in biennial caraway (Arganosa et al. 1998). This can also vary according to the genotype, seeding date, and location. Our DH lines show a lower carvone content, however the parental lines were also similar. The industry prefers caraway oil to have a higher carvone content (50–60%) (Weiss 2002) than was observed in any lines tested in this trial, including the parental lines. Carvone content increases as the seed ripens, and consequently, biennial types of caraway tend to have a higher carvone percentage than the annual types used in this project.

Fennel

In both 2003 and 2004, seed germination of the DH lines of oilseed fennel was quite variable. In 2003, four of the 14 lines were non-viable, while in 2004, five of the 11 lines tested did not germinate.

Of the 16 viable DH lines (10 lines from 2003 and 6 lines from 2004), five had 100% germination while the germination percentages of the other lines varied from 25 to 60% (data not shown). The parental line consistently had a high germination percentage (95–100%). There was significant variation from line to line in plant height of the fennel at the end of the growing season (ca 120 days after planting). In 2003, two of the DH fennel lines (DH-37 and DH-90) were approximately twice the height of the parental line (60 cm; Fig. 2a). The heights of the DH lines tested in 2004 more closely resembled the parental lines (Fig. 2b). Within the DH lines, plant heights were more uniform than the parental material, likely reflecting the genetic homogeneity characteristic of DH. Aside from differences in height, the morphology and rate of development of the DH fennel lines were similar to the parental line. In both years, the growing season was too short to allow development of seed in any of the fennel lines tested.

Conclusion
The Apiaceae species have been considered recalcitrant when it comes to androgenesis, especially microspore culture. Very little research has been published in this area and there were no efficient protocols available for generating doubled haploids prior to the methods developed by Ferrie et al. (2005, 2010). The objective of this project was to compare the agronomic performance of DH lines of dill, caraway, and fennel relative to the parental lines, while also surveying the DH material for traits that could be useful in subsequent crop improvement. Field comparisons of DH lines and their parental controls have not been previously reported in the Apiaceae, but there are reports in the Brassica species and barley (Hordeum vulgare) (Park et al. 1976; Friedt and Foroughi-Wehr 1983; Powell et al. 1986). In those species differences were observed between the DH lines and the parental line for a number of characteristics (e.g. yield, height, maturity). Generally, the DH lines were inferior (Palmer et al. 1996), but superior lines could be identified (Friedt and Foroughi-Wehr 1983). In our study, a significant proportion of the DH lines of dill, caraway, and fennel proved to have limited seed viability. Generally, this has not been observed in DH lines of other species. However, many of the Apiaceae DH lines that were viable showed promise in field trials. DH lines of dill were identified that were earlier maturing, more uniform in stature, and higher yielding than the parental lines. A single DH line of caraway with exceptional yield potential was also identified. Seed oil content and composition of the DH lines of dill and caraway were generally comparable to the parental lines, suggesting that the improvement in field performance was achieved without compromising crop quality. Our results show that Apiaceae doubled haploidy techniques can generate lines that have the potential for commercial production as well as lines that can be incorporated into a breeding program. Further efforts to create and characterize DH lines of dill and caraway with superior agronomic or quality characteristics are ongoing.

Acknowledgments

The authors thank Jackie Bantle and William Hrycan for their technical assistance.

Financial assistance from Saskatchewan Agriculture, Food and Rural Revitalization, Agriculture Development Fund was greatly appreciated.

References

Arganosa GC, Sosulski FW, Slinkard AE (1998) Seed yields and essential oils of annual and biennial caraway (Carum carvi L.) grown in Western Canada. J Herbs Spices Med Plants 6:9–17

Bailer J, Aichinger T, Hackl G, de Hueber K, Dachler M (2001) Essential oil content and composition in commercially available dill cultivars in comparison to caraway. Ind Crop Prod 14: 229–239

Beveridge JL, Dalziel J, Duncan HJ (1981) The assessment of some volatile organic compounds as sprout suppressants for ware and seed potatoes. Potato Res 24:61–76

Ferrie AMR, Bethune T, Kernan Z (2005) An overview of preliminary studies on the development of doubled haploid protocols for nutraceutical species. Acta Physiol Plant 27: 735–741

Ferrie AMR, Bethune TD, Mykytyshyn M (2010) Microspore embryogenesis in Apiaceae. Plant Cell Tissue Organ Cult. doi: 10.1007/s11240-010-9770-0

French DH (1971) Ethnobotany of the Umbelliferae. In: Heywood VH (ed) The biology and chemistry of the Umbelliferae. Linnean Society of London, Academic Press, London, pp 385–412

Friedt W, Foroughi-Wehr B (1983) Field performance of androgenetic doubled haploid spring barley from F1 hybrids. Z Pflanzenzuchtg 90:177–184

Helander IM, Alakomi HL, Latva-Kala K, Mattila-Sandholm T, Pol I, Smid EJ, Gorris LGM, von Wright A (1998) Characterization of the action of selected essential oil components on Gram-negative bacteria. J Agric Food Chem 46:3590–3595

Lee SK, Tsao R, Peterson C, Coats JR (1997) Insecticidal activity of monoterpenoids to western corn rootworm (Coleoptera: Chrysomelidae), two spotted spider mite (Acari: Tetranychidae), and house fly (Diptera: Muscidae). J Econ Entomol 90:883–892

Oosterhaven K, Leitao AC, Gorris LGM, Smid EJ (1996) Inhibition of potato sprout by carvone enantiomers and their bioconversion in sprouts. Potato Res 38:219–230

Palmer CE, Keller WA, Arnison PG (1996) Utilization of Brassica haploids. In: Jain SM, Sopory SK, Veilleux RE (eds) In vitro haploid production in higher plants. Kluwer, Dordrecht, pp 173–192

Park SJ, Walsh EJ, Reinbergs E, Song LSP, Kasha KJ (1976) Field performance of doubled haploid barley lines in comparison with lines developed by the pedigree and single seed descent methods. Can J Plant Sci 56:467–474

Peirce A (1999) The American pharmaceutical association practical guide to natural medicines. Stonesong Press, New York

Powell W, Caligari PDS, Dunwell JM (1986) Field performance of lines derived from haploid and diploid tissues of Hordeum vulgare. Theor Appl Genet 72:458–465

Salom SM, Gray JA, Alford AR, Mulesky M, Fettig CJ, Woods SA (1996) Evaluation of natural products as antifeedants for the pales weevil (Coleoptera: Curculionidae) as a fungitoxins for leptographium procerum. J Entomol Sci 31:453–465

Smid EJ, de Witte Y, Gorris LGM (1995) Secondary plant metabolites as control agents of postharvest penicillium rot on tulip bulbs. Postharvest Biol Technol 6:303–312

Thomas WTB, Forster BP, Gertsson B (2003) Doubled haploids in breeding. In: Maluszynski M, Kasha KJ, Forster BP, Szarejko I (eds)
Doubled haploid production in crop plants, a manual. Kluwer, Dordrecht, pp 337–349

Weiss EA (2002) Spice crops. CABI Publishing, Wallingford Plant Cell Tiss Organ Cult (2011) 104:407–413 413

you have reached the end of Objective 4 - Click here to return to the top of the page

Another Website by Heartland Graphics   Copyright HSSA: Herb, Spice and Specialty Agriculture