• cypsela burial depth;
  • dormancy;
  • emergence;
  • Scotch thistle;
  • soil microenvironment


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • 1
    Cypselas (fruits) of Scotch thistle (Onopordum acanthium) germinate unpredictably over a long period of time. We evaluated the effects of soil type, burial depth and collection date on the emergence patterns of two populations of O. acanthium, ESW and Quarry (Q).
  • 2
    From each of four separate collections made throughout 1996, five replicates of 200 cypselas each were placed on the surface or buried at depths of 3 or 15 cm in both sand and silt-loam soils (i.e. field conditions) and emergence was recorded biweekly between August 1996 and September 1999.
  • 3
    Emergence was intermittent over the 3 years, with a higher total percentage from silt loam (17%) than from sand (9%), and from 3 cm depth (18%) compared with the surface (8%).
  • 4
    After 3 years, all remaining viable, split (germinated but not emerged), decayed and unfilled cypselas were retrieved. Total germination over the 3 years (emerged plus split) was up to 77% at 3 cm in silt loam but always less than 10% at 15 cm in both soils for both populations. Although a higher percentage of cypselas germinated under controlled conditions when retrieved from loam compared with sand (89% vs. 66%), a significant number from all treatments did not germinate until scarified.
  • 5
    Cypselas that ripened under the warmest temperature (collection 3) had the highest emergence in the year of collection (1996), while more of those that ripened under cooler temperatures (collections 1 and 4) emerged in later years (1998 and 1999). Both initial and later emergence patterns for both populations varied greatly with collection date. This may be the first time that seedling emergence from the seed bank for different collections has been monitored over several years.
  • 6
    We compared waxes, lignins and soluble phenolics in the coats of cypselas retrieved from soil (early germinators) with those that germinated only after scarification. More surface wax was found in dormant cypselas.
  • 7
    Understanding these complex germination/emergence patterns may help develop control policies for O. acanthium.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

One of the most important attributes of many weed, ruderal or colonizing species is the production of seeds with varying degrees of dormancy (Baskin & Baskin 1998; Cavers et al. 2000). Such species characteristically possess large seed banks (Cavers & Benoit 1989), and their viable seeds germinate intermittently over months or even years (Salisbury 1961). Control of weeds with such behaviour is difficult because new seedlings emerge soon after each episode of weed destruction (Harper 1977).

Variation in seed dormancy can be attributed to variation in genotypes within the species (Baskin & Baskin 1998) or within a single population (Cavers & Harper 1966). Variation in factors extrinsic to the seed, such as the nature of the substrate on which it falls (Fenner 1991; Gutterman 2000) and depth in the soil (Taylorson 1970; Baskin & Baskin 1998), can also have a great influence on the timing of germination. In fact, seeds from the same plant may show very different germination patterns after overwintering on the soil surface vs. beneath the soil (Cavers et al. 2000). One factor that has not been examined in detail for species that ripen seeds over a period of months is the date of maturation. Seeds ripened at different times and under different temperatures, day lengths, moisture levels and stages of maturity of the parent plant have been shown to differ in germination (Baskin & Baskin 1998), but they have not been compared for performance in the seed bank.

Germination of individual seeds may be governed by either or both of embryo or seed coat dormancy (Bradbeer 1988; Bewley & Black 1994). Coat-imposed dormancy can be caused by seed coat hardness (e.g. the degree of lignification), by the presence of soluble phenolics in the seed coat or by the deposition of waxes on the surfaces of seeds, which either act as water repellents or prevent germination inhibitors from leaching out (Williams & Hoagland 1982; Stabell et al. 1996; Werker 1997). However, once in the seed bank, environmental factors may cause changes in the conditions that impose dormancy (Fenner 1985; Blackshaw 1992).

Scotch thistle (Onopordum acanthium L.), a carduine thistle (Asteraceae), is usually a monocarpic biennial, but under certain conditions can be an annual or a short-lived perennial, reproducing almost entirely by cypselas. A mature plant can grow up to 3 m in height. In London, Ontario, O. acanthium flowers from late June to early October, and ovule fertilization occurs by self- and cross-pollination. Depending on size, a plant may produce as few as 100 or as many as 50 000 cypselas (Qaderi 1998). Onopordum acanthium is found in many parts of the world, in habitats from disturbed areas to agricultural fields, and can cause problems in infested areas (Qaderi 1998; Smith et al. 1999). This noxious weed has an unusual pattern of cypsela population dynamics, due to its strongly intermittent germination. Some cypselas germinate shortly after dispersal, while others remain dormant for several months to many years, so that, in long-term seed bank studies, it has proved to be one of the longest-lived weed species. For example, in the Duvel buried seed experiment (Toole & Brown 1946) up to 46% of O. acanthium cypselas germinated after 39 years and, in contrast to most other species, germination percentage actually increased with increasing duration of storage. Following cultivation, plants can appear on areas where there has been no seeding for several years (Parsons 1973). Roberts & Chancellor (1979) concluded that the level of innate dormancy in their sample of O. acanthium appeared to be somewhat greater than that in the other species they tested. They found that O. acanthium not only differed in emergence pattern from other species, but its cypselas also appeared to have a greater capacity for persistence in cultivated soil. Cavers et al. (1995) have also reported that cypselas of Onopordum sp. (O. acanthium, O. illyricum L. or a hybrid between them) vary greatly in dormancy, even within the same population. Information about the seed bank of O. acanthium is therefore important for determining effective control methods.

We hypothesized that variation in the degree of dormancy is influenced by the microenvironment of the seed, both before and after dispersal. We designed an experiment to investigate the effects of soil type, burial depth and cypsela maturation date on germination/emergence patterns, changes in dormancy and retention of viability in O. acanthium cypselas over a 3-year period. We also measured some chemical constituents (waxes, lignins, soluble phenolics) from three categories of cypselas (split coats remaining in the soil, coats from cypselas germinated in the incubator and non-germinated but viable cypselas) to investigate a possible correlation between variation in chemical properties and intermittent germination.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References


Bulk collections of O. acanthium cypselas were taken four times (5 August, 20 August, 4 September and 19 September 1996) from 100 randomly selected plants of each of two populations: one at the Environmental Sciences Western (ESW) research station and another at the Clarke Sideroad Quarry (Q) in London, Ontario. For further details of the populations see Table 1. The cypselas were cleaned manually (pappi removed) with the aid of a seed blower and aborted cypselas were discarded.

Table 1.  Description of collection sites of O. acanthium populations in London, Canada
PopulationLatitude and longitudeSoil textureSoil pHHabitat
ESW (Environmental Sciences Western research station)43°04′25′′ N 81°20′11′′ WSilt loam7.5Arable field, uncultivated for one year. This population originated in 1992 from gravel soil (pH 7.9), woodland edge 20 m east of Thames River, frequently flooded habitat (42°59′ 52″ N, 81°16′ 05″ W)
Q (Quarry, Clarke Sideroad)43°02′25′′ N 81°11′38′′ WLight gravelly loam8.030–50 m from edge of gravel pit, open area amongst deciduous trees, 20–40 m from Thames River, never flooded


Within 3 days of each collection, cypselas were placed on the surface, or buried at 3 or 15 cm depths in each of sand and silt-loam soils in an open, level site at the ESW research station (five replicates of 200 cypselas per treatment). Each replicate was placed inside an inverted, 2 L circular plastic drinks bottle (10.5 cm in diameter) that had had both ends removed before burial, with approximately 3 cm above the soil surface. At each burial date, six additional bottles were inserted into the soil as controls to check if any cypselas of O. acanthium remained in the soil from previous years. On each soil, the bottles were arranged in a completely randomized design and a fibreglass mesh screen was spread on the top of the entire set to prevent birds feeding on cypselas. Mean monthly maximum and minimum air temperatures and total precipitation, recorded at London International Airport (< 3 km from site Q), are shown in Fig. 1. Emerged seedlings with fully expanded cotyledons were counted and removed biweekly from August 1996 to September 1999. Cypselas from the surface that produced a radicle (2 mm long or more) and then died (from desiccation) were counted as emerged.


Figure 1. Mean monthly maximum and minimum temperatures and total monthly precipitation in London, Ontario from August 1996 to September 1999. Data taken from Environment Canada, Ontario Region, London Airport, 43°01′ 52′′ N, 81°09′20′′ W (Environment Canada 1996, 1997, 1998, 1999).

Download figure to PowerPoint


After 3 years, non-germinated cypselas were exhumed from the soils and extracted using a soil flotation method (modified from Malone 1967) as follows: the soil sample (all soil inside a bottle) containing cypselas was dispersed in an aqueous solution of sodium hexametaphosphate (10 g 200 mL−1), sodium bicarbonate (5 g 200 mL−1) and magnesium sulphate (25 g 200 mL−1). Each soil sample was mixed thoroughly with 200 mL of the prepared solution in a 500-mL Erlenmeyer flask. The flask was capped with a glass stopper and shaken vigorously. The organic matter was decanted through 1.18-mm and 850-µm sieves (stacked). Cypselas and split cypsela coats remaining on the sieves were collected and spread on a piece of paper at room temperature until dry. Sound (apparently viable) cypselas were then separated from the remainder (split, empty, decayed) and used for germination experiments. Those cypselas that had no embryos, but with the two halves of the cypsela coat remaining attached, which probably arose from cypselas that had germinated but not emerged, were classed as split. The ‘split’ cypsela coats remained attached at both ends, thus enabling them to be distinguishable from the coats of emerged seedlings. Cypselas with somewhat soft coats (which separated with light pressure from forceps) and no embryo were designated as empty. Cypselas with degraded embryos and coats were classed as decayed.


Sound cypselas from each of the five burial replicates were placed in a 9-cm diameter glass Petri dish on one layer of blue germination filter paper (Anchor Paper Co., St Paul, Minnesota, USA) initially moistened with 10 mL of distilled water. More water was added each day as needed. The cypselas were set to germinate in an incubator at 25 °C for 14 h in the light, and at 10 °C for 10 h in the dark. Light was provided by two cool white fluorescent tubes (mean photosynthetically active radiation (PAR) of 39.4 µmol photons m−2 s−1 at the level of the Petri dishes, n= 20) situated c. 25 cm above the surfaces of the Petri dishes. Germinated cypselas (radicle 2 mm or longer) were counted daily, removed from the Petri dishes and kept for chemical analyses. The experiment was terminated after a 5-day period with no germination (typically after 45 days). Remaining firm non-germinated cypselas were cut 1 mm from the cotyledonary end and returned to the incubator for 15 days to distinguish the non-germinated viable cypselas from the non-germinated dead ones. Coats from the incubator-germinated and non-germinated cypselas were also collected for chemical analyses.


Wax components

Total surface waxes were removed (three replicates) by brief immersion (30 s) of 0.07–5.33 g of cypsela coats in 80 mL of chloroform (CHCl3) at room temperature. The CHCl3 extract was filtered, concentrated by rotary evaporation and the residue weighed. The wax residue was re-dissolved in dichloromethane (CH2Cl2) and transferred into a 4-mL vial. Next, 0.02 mg tetracosane (Sigma-Aldrich) was added as an internal standard and the solution was dried under a stream of nitrogen, re-dissolved in pyridine (60 µL) and derivatized with 60 µL Bis- trimethylsilyl-trifluoroacetamide (BSTFA; 99% with 1% trimethylchlorosilane) for 40 min at 70 °C in a water bath. The derivatized wax residue was diluted with CH2Cl2 (40 µL) and 1 µL injected into a Varian GC-MS (gas chromatography-mass spectrometer) fitted with a capillary column (CP-Sil 5 CB Low Bleed/MS, 30 m × 0.25 mm ID, DF = 0.25 µm; Chrompack). Components were eluted with the following programmed temperature gradient: 2 min isothermal heating at 50 °C followed by a 40°C min−1 oven ramp to 200 °C, a 3 °C min−1 oven ramp to 300 °C and isothermal heating at 300 °C until the end of the 50 min run (approx. 9 min). The injector and detector temperatures were set at 250 °C. Helium was used as a carrier gas at 1 mL min−1 (83 kPa column head pressure). Electron impact ionization mass spectra were recorded with a Saturn 2000 mass detector (Varian) using a scanning range of 40–650 m/z. The identity of each compound was determined based on its retention time, mass spectrum and comparison with authentic standards (Sigma).

Soluble phenolics

Soluble phenolic compounds were Soxhlet-extracted (three replicates) from the wax-extracted cypsela coats, using methanol as follows: cypsela coats were ground with a pestle in a mortar in liquid nitrogen and c. 1 g of the homogenized tissue from each sample was weighed into a micro-Soxhlet thimble. The residue was extracted for 8 h with 80 mL methanol and then for another 8 h with 80 mL of a mixture of chloroform-methanol (3 : 1). The thimble containing the sample was dried at room temperature for 24 h and weighed. Extracts from MeOH and MeOH-CHCl3 were concentrated by rotary evaporation, quantitatively transferred into an 8-mL vial with MeOH, and dried under reduced pressure. After re-dissolving in MeOH (1 mL g−1 dry residue) and centrifuging (17 300 g, 5 min), the extracts were transferred into 2 mL HPLC vials and analysed by reversed phase, analytical HPLC. Samples (20 µL) were injected into a Nucleosil C-18 column (4.6 × 150 mm, 5 µm, 100 Å) via an autosampler and eluted with a solvent gradient of water-acetonitrile-methanol (1 : 1 : 1, solvent B) in 1.5% aqueous phosphoric acid (solvent A) as follows: 10% to 25% over 5 min followed by 25% to 75% over 25 min at 1 mL min−1. The column was washed with 100% solvent B and re-equilibrated with 10% B in A for 15 min between samples. Compounds were detected at 240 nm.

Lignin content

The extractive-free cypsela coat residues were used for lignin analysis by the DFRC (derivatization followed by reductive cleavage) method essentially as described (Lu & Ralph 1997). The dried filtrate after zinc cleavage was acetylated in 1.5 mL of CH2Cl2 (containing acetic anhydride 0.2 mL and pyridine 0.2 mL) for 40 min. The acetylated residue was coevaporated with ethanol three times under reduced pressure to remove all volatile compounds. The final derivatized sample was dissolved in 0.8 mL of CH2Cl2, and a 1-µL aliquot injected into a GC-MS, fitted with a capillary column (CP-Sil 5 CB Low Bleed/MS, 30 m × 0.25 mm ID, DF = 0.25 µm; Chrompack). Compounds were eluted with the following programmed temperature gradient: 1 min isothermal heating at 140 °C followed by a 3 °C min−1 oven ramp to 240 °C and isothermal heating at 240 °C until the end of the 40-min run (approx. 5.7 min). The injector and detector temperatures were set at 250 °C. Helium was used as a carrier gas at 1 mL min−1 (83 kPa column head pressure). Electron impact ionization mass spectra were recorded with a Saturn 2000 mass detector (Varian) using a scanning range of 40–650 m/z. The amount of each lignin monomer was determined based on the internal standard (tetracosane) and the response factor of the standard.


Final emergence percentages in the soil, final germination percentages in the incubator, total numbers of split cypselas from the field, total numbers of viable but non-germinated cypselas in the incubator and total numbers of non-viable cypselas were each analysed by means of a balanced anova on both populations together and on each of the populations separately (McKenzie et al. 1995), after an arc-sine square root transformation to normalize the variance (Zar 1999). A Tukey's multiple comparison test was then used to determine differences between treatments (SAS 1989). A Pearson correlation analysis was performed to determine the relationship between different cypsela categories. Data from the wax, lignin and phenolic analyses were changed first to percentages and then analysed by a balanced anova (Sokal & Rohlf 1995).


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References


Some seedlings emerged from almost all collections on both soil types for both the surface and 3 cm depth treatments during each year from 1996 to 1999. However, half of the population × depth × soil combinations for the last collection (collection 4) showed no germination in 1996 and the last two collections from the Q population showed no germination from the 3 cm depth in sand during 1998 (Table 2). No cypsela emerged from the 15 cm depth treatment nor from the control plots. In every year, emergence was intermittent (Table 3), with new seedlings appearing as early as March (1998) and as late as October (1996 and 1997).

Table 2.  Total number of emerged seedlings (per depth per collection) from two O. acanthium populations placed on the surface (0 cm) or buried at 3 cm depth from August 1996 to September 1999 in London, Canada
Soil typeSoil depthCollection numberTotal emergence (1996–1999) (n = 1000)Total number of emerged seedlings out of 1000Total percentage emergence per depth
Sand0 cm1141337 
 Subtotal (n = 4000)19363952964.8
 3 cm11842154244 
 Subtotal (n = 4000)58496424412314.6
Loam0 cm1995353227 
 Subtotal (n = 4000)35914139981089.0
 3 cm133011708673 
 Subtotal (n = 4000)10947156730015629.3
Sand0 cm1551221121 
 Subtotal (n = 4000)23799075635.9
 3 cm18577035 
 Subtotal (n = 4000)44518724041411.1
Loam0 cm11309802615 
 Subtotal (n = 4000)483472491177012.1
 3 cm11568864319 
 Subtotal (n = 4000)7891024511993719.7
Total (n = 32 000)4184532219988656713.1
Table 3.  Emergence (total/month) of two populations (ESW, Q) of O. acanthium cypselas placed on the surface (0) or buried at 3 cm (3) depth in sand (S) and silt-loam (L) soils. Emergence data recorded from 1996 to 1999 are based on combined data from five replicates of 200 cypselas, for each of four different collection dates, ntotal= 4000. Only months in which emergence was recorded are included. A dash (–) indicates no emergence in that month
YearMonthTotal emergence
ESW S-0ESW S-3ESW L-0ESW L-3Q S-0Q S-3Q L-0Q L-3

There was no difference in seedling emergence between the two populations but all other main effects and four two- and three-way interactions were significant (Table 4). A higher total percentage emerged from silt loam (17%) than from sand (9%) (F = 9.54, P= 0.002) and from 3 cm depth (18%) than from the surface (8%) (Table 2). Overall, emergence was concentrated in the months of April and May (1997–99) or in September (1996), when relatively high soil moisture contents were accompanied by large temperature fluctuations (Fig. 1 and Table 3). However, for both populations large numbers of cypselas on the sand surface emerged in late summer or early autumn in 1997, 1998 and 1999 (Table 3). These emerged seedlings made up over 45% of seedlings from this treatment, compared with negligible numbers of late summer/autumn seedlings from buried cypselas or cypselas on loam. When the data for all years were combined, the highest (1094) and the lowest (193) total numbers of emerged seedlings were both from the ESW population (respectively from 3 cm in silt loam and on the surface of sand, Table 2). For individual collections over the entire period of the experiment, the highest (336) and the lowest (14) numbers of emerged seedlings were again recorded for the ESW population, from the second collection buried at 3 cm in silt loam and from the first collection of cypselas placed on the sand surface, respectively (Table 2).

Table 4.  Analysis of variance for population, soil type, burial depth and collection date effects on field emerged, field splits, incubator-germinated, non-germinated but viable and non-germinated dead cypselas of Scotch thistle. Significance values: *P < 0.05; **P < 0.01; ***P < 0.001
Sourced.f.Field emergedField splitsIncubator-germinatedNon-germinated viableNon-germinated dead
Population (P)116.60.713.80.4560.44.5*369.
Soil type (ST)11247.954.5***5000.5152.9***1052.88.4**13722.0141.9***1259.039.4***
Burial depth (BD)212104.7528.6***9899.3302.7***17134.8136.9***6857.870.9***705.922.1***
Collection date (CD)374.73.3*214.46.6***134.81.1663.16.9***2650.482.9***
P × ST12.**14.40.4
P × BD2246.810.8***121.83.7*403.83.2*133.41.436.01.1
P × CD317.80.834.*
ST × BD2327.214.3***809.424.7***6457.751.6***1398.914.5***173.45.4**
ST × CD365.32.8*13.20.4279.52.2370.33.8*160.75.0**
BD × CD637.11.6150.64.6***150.51.2253.52.6*65.42.0
P × ST × BD223.
P × ST × CD38.
P × BD × CD662.62.7*
ST × BD × CD630.31.342.71.3378.33.0**
P × ST × BD × CD67.00.318.20.6120.
Error19222.9 32.7 125.1 96.7 32.0 

Large differences were recorded among cypselas from different collection dates. In O. acanthium, the critical period of cypsela maturation, during which temperature influences dormancy, occurs in the 2 weeks prior to maturity. For the 3 cm depth, initial germination (i.e. by autumn 1996) was greater for cypselas completing maturation under warmer temperatures (350 seedlings from collection 3) rather than cooler temperatures (a total of 30 from collections 1 and 4) (Fig. 2 and Table 2). In contrast, in 1998 and 1999, 257 and 279 seedlings emerged from this depth for collections 1 and 4, respectively, but only 73 for collection 3 (Table 2).


Figure 2. Mean daily maximum and minimum temperatures (lines) and total daily precipitation (bars) in London, Ontario from 1 July to 31 October 1996. The lower four graphs show total emergence per day from surface and 3 cm depth for each of four collections (C1-C4) of the two populations, ESW (□) and Q (▪), in 1996. Vertical arrows indicate the collection dates.

Download figure to PowerPoint



Split cypselas were the remains of cypselas that had germinated in the soil but died without emerging. A strong positive correlation was found between the number of splits and the number of emerged seedlings (r = 0.826) (Figs 3 and 4).


Figure 3. Fate of O. acanthium cypselas from the ESW population placed on the surface or buried at 3 cm or 15 cm depths in sand and silt-loam soils for 3 years (August 1996 to September 1999). Proportions of emerged (□), field splits (bsl00006), incubator-germinated (▧), non-germinated but viable (▪) and non-viable (▥) cypselas are shown for each of the four collection dates separately.

Download figure to PowerPoint


Figure 4. Fate of O. acanthium cypselas from the Q population (symbols as in Fig. 3).

Download figure to PowerPoint

The greatest number of splits was retrieved from 3 cm and the fewest from 15 cm (F = 151.54, P < 0.001). There were fewer from sand than from silt loam (F = 39.30, P < 0.001). Differences between all main effects except population and all two-way interactions involving burial depth were significant (Table 4). The numbers of cypselas retrieved as splits were highest from the surface and 3 cm in silt loam. The first collection date had the fewest splits. For the second collection, there were large differences among burial depths, with the most from 3 cm and the fewest from 15 cm. The percentage of splits retrieved from 15 cm depth was only 0.5% for collections 1 and 2 but 3–6% for collection 4 (Figs 3 and 4).

Germination in the incubator

For the cypselas that germinated readily after retrieval, significant differences were recorded: between populations, with more cypselas from Q germinating than from ESW; between soil types, with 89% of apparently viable cypselas from silt loam germinating vs. only 66% from the sand; and among burial depths, with the highest numbers that germinated coming from 15 cm (Table 4 and Figs 3 and 4). Several interactions were also significant (Table 4). Cypselas from 15 cm in silt loam had the highest incubator-germination percentage, while those from 3 cm of the same soil type had the lowest. The greatest differences among collections were for the cypselas from the surface of sand, where collection 3 showed greatest emergence (Figs 3 and 4). Cypsela germination in the incubator was negatively correlated with both field emergence (r = −0.608) and number of field splits (r = −0.605).

Viable but non-germinated in the incubator

For viable cypselas that did not germinate in the incubator (i.e. those that only germinated after scarification), all main effects except population were significant (Table 4), with increased proportions from sand, especially for the ESW population, and from collection 1 rather than collection 4. However, the vast majority of the resulting seedlings had little vigour, and were stunted or dwarfed. Several two-way interactions were also significant (Table 4). The number of non-germinated but viable cypselas was negatively correlated with the number that germinated in the incubator (r = −0.571).

Non-viable cypselas

Up to 30% of the cypselas that did not germinate in the field or the incubator were non-viable (Figs 3 and 4). These included cypselas that were empty (but not split) or decayed when retrieved from the soil, plus those that did not germinate after scarification. Differences in numbers of non-viable cypselas between soil types, burial depths and collection dates were significant (Table 4). More non-viable cypselas were retrieved from the sand and from the surface treatment (Figs 3 and 4) and for the fourth collection (vs. collections 1 and 2, F= 54.89, P < 0.001). Several two-way interactions were also significant (Table 4). The number of non-viable cypselas was not different between populations but was higher for population Q in collection 3 and for ESW in collection 4. For all except the third collection, more non-viable cypselas were found from sand than from silt loam. A negative correlation was found between the numbers of non-viable cypselas and the numbers that germinated in the incubator (r = −0.413).


Onopordum acanthium cypsela coats contain a predominantly syringyl lignin, with only traces of guaiacyl units. A lower, but not significantly lower, lignin content was found in non-germinated (but viable) cypselas than in either field- or incubator-germinated ones (Table 5).

Table 5.  Lignin content (µg.g−1 dry weight) from coats of O. acanthium cypselas retrieved from sand and silt-loam soils. Cypselas were categorized as non-germinated (but viable), splits (i.e. field-germinated but not emerged) or incubator-germinated. Means followed by different letters are significantly different (P < 0.05) according to Tukey's multiple comparison test
SoilCypsela categoryBurial depth (cm)Pooled
SandNon-germinated40.26 ± 2.97b42.13 ± 8.76ab41.46 ± 5.09b 
 Field splits60.59 ± 4.66ab49.58 ± 2.56ab71.66 ± 9.53ab 
 Incubator-germinated70.45 ± 6.67a
LoamIncubator-germinated56.34 ± 4.82ab

A significantly higher surface wax content was obtained from non-germinated cypselas than from either field- or incubator-germinated ones (F = 272.41, P < 0.001). In particular, total surface wax (mg g−1 dry weight) was higher for non-germinated but viable cypselas from sand (1.92 ± 0.06) than for field- (0.1 ± 0.04) or incubator-germinated cypselas (0.28 ± 0.07) from the same soil type or for incubator-germinated ones from loam (0.16 ± 0.01). However, no statistically significant difference was found among the latter three. The surface wax of O. acanthium is composed of long-chain alkanes, fatty acids and alcohols, but predominantly C-16 and C-18 fatty acids. The number and the amount of wax components were found to be higher for the non-germinated cypselas than for the other categories of cypselas extracted from the soil (Table 6).

Table 6.  Surface wax components (relative amount/1 mg cypsela) from O. acanthium cypselas of the ESW population placed on the surface or buried at 3 or 15 cm depths in sand and silt-loam (only incubator) soils in 1996 and retrieved in 1999 in London, Canada. For each wax component, the treatment containing the lowest measurable amount was set at 1.0, with the amounts of all others calculated relative to it. Data are directly comparable within rows, but not between rows. Trace = minor peak (not integrated); −= none detected. In the first column one = alkane, oate = fatty acid and ol = alccial.
Wax componentsNon-germinated but viableRetrieved as splits from soilGerminated in incubator
Sand, 0 cmSand, 3 cmSand, 15 cmSand, 0 cmSand, 3 cmSand, 15 cmSandLoam
C16oate (16 : 0)11.327.423.
C16oate (16 : 1)21.334.453.41.0Trace5.513.511.5
C18oate (18 : 0)192.4692.7376.
C18oate (18 : 1)

Overall, there were only small differences in the quantity and quality of individual components, as evidenced by HPLC analyses, in the methanol extractives obtained from cypsela coat walls retrieved from the different burial treatments (data not shown).


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This study provides an exceptionally thorough record of the fates of propagules from a species that ripens and releases cypselas over a prolonged period of the growing season. The cypselas were left on or in the soil, not enclosed in bags or containers, and allowed to germinate and emerge naturally over the 3-year period of study. After 3 years it was possible to separate the remaining cypselas into those that could have germinated and those that had died without germinating or had never been viable. Consequently, there are many differences between our experimental approach and these used in other studies where seeds were stored in the field from collections made at different times (e.g. Baskin & Baskin 1995).


Milberg & Andersson (1998) emphasized that dormancy levels in seed batches can vary within a species and that ‘intraspecific variability in dormancy persists even after seeds have been stratified in soil’. In a recent paper, Qaderi & Cavers (2000a) demonstrated different germination responses from several local populations of O. acanthium, including the two used here, but differences in their dormancy in this experiment were less.


Many experimental studies have shown a correlation between the depth of seeds in the soil and emergence patterns in the field (Harper 1977; Redmann & Qi 1992; Grundy & Mead 1998).

Viable, non-germinated seeds at the soil surface in natural areas can be exposed to predation and are destroyed to a large extent (Parker et al. 1989). Although Baker (1989) concluded that germination usually is enhanced at the soil surface, he and Fenner (1985) cited several studies where seeds on the surface were inhibited from germinating compared with lightly buried seeds. This was also the case in this experiment, where there was more emergence from 3 cm than from surface-stored cypselas. Protective mesh prevented predation by birds and small mammals but otherwise the cypselas were exposed to natural conditions. One feature of this study was the lack of litter accumulation over the cypselas, and therefore a lack of burial, because the experiment was located in an open site surrounded by either bare soil or frequently mown grass.

We have not found any other multiyear study comparing seed germination, mortality and retention of viability for surface-lying vs. lightly buried seeds. Many non-germinated cypselas did lose vigour over the 3-year period, especially on sand, but there were substantial numbers (> 20%) that produced vigorous seedlings after 3 years. As for a variety of weed species, loss in vigour was greatest on the surface and least at depths around 15 cm (Roberts 1981), but O. acanthium had much higher percentage survival on the surface than has been recorded over shorter periods for other species.

As seeds on the soil surface experience the most wetting and drying, they can acquire secondary (induced) dormancy (Doucet & Cavers 1996) or may germinate only when sufficient water is available. In addition, less emergence from the soil surface may have reflected the fact that cypselas on the surface were exposed to extremely high summer temperatures (> 35 °C) and low winter temperatures (down to −24 °C) (Environment Canada 1997, 1999), which could have enhanced or induced their dormancy (Baskin & Baskin 1998).

The highest emergence numbers were obtained from cypselas buried 3 cm deep (Tables 2 and 3) and the largest number of splits was also obtained from that depth (Figs 3 and 4). The seed bank remaining after 3 years was therefore smallest at this depth, particularly on loam soil.

Lack of germination of cypselas at the 15 cm depth might have been related to poor germination conditions, such as darkness, lack of temperature fluctuation, and poor gas exchange (CO2 and O2) (Baskin & Baskin 1998). However, as the number of non-viable cypselas was also lowest there (Figs 3 and 4), it may be that conditions that imposed quiescence might also have favoured the retention of viability and thus long-term survival of the species. Roberts & Feast (1972) also found much better survival of seeds of many species at 15 cm than at shallower depths, but much more seedling emergence from shallow depths. Cypselas from deep in the soil can be brought to the surface by cultivation (Cavers & Benoit 1989; Mohler 1993), exposing propagules that are viable and poised to germinate as soon as environmental conditions are right.


In addition to the effects of depth, the microenvironment varies with soil type, and seeds of different species have different ranges of tolerance (Cavers & Benoit 1989; Parker et al. 1989). Soil moisture, aeration, fertility, organic content and biological activity all affect populations of each species in the seed bank. There were large differences in emergence between cypselas buried in sand and those buried at the same depths in or on silt loam, with loamy soil and deep planting creating favourable environments for the retention of dormancy and viability of O. acanthium cypselas.


Onopordum acanthium is one of many species in which individual plants flower and ripen seeds over several months of the growing season. During this period, both predictable changes (e.g. physiological age, day length) and unpredictable ones (e.g. temperature, soil moisture) can affect seed formation.

Baskin & Baskin (1998) summarized published results on environmental factors that cause preconditioning effects on seeds and thus ultimately affect germination. Attributes such as seed colour, dormancy and rapidity of germination all changed with differences in age of the parent.

There were large differences in emergence patterns among cypselas from different collection dates. Few cypselas from collections 1 and 4 had emerged in the first autumn, but many emerged in the third and fourth years, especially from the 3 cm depth in silt loam. In contrast, 72% of all seedlings that emerged in the first autumn (1996) came from collection 3, whose cypselas ripened under the preceding warmer conditions (Fig. 2). Qaderi & Cavers (2002) demonstrated that cypselas of O. acanthium ripening in a cool year in the field were less strongly dormant than those of the same population ripening in a warmer year. In the second year (1997) the largest proportion of emerged seedlings (34%) came from collection 2. These different results from the four collections contributed strongly to the overall intermittent pattern of emergence shown in Table 3.

For all treatments, the number of non-viable cypselas increased significantly from the first to the last collection, suggesting that cypselas maturing later in the season contribute less to reproduction in this species than those that mature earlier. This conclusion is further supported by the great disparity in numbers of cypselas produced over the season, with the largest number ripening in the large, early capitula.


Even though the outer structural characteristics that affect seed impermeability are genetically controlled, environmental conditions can have a marked influence as well (Priestley 1986). Seed/fruit coat hardness and thickness can often impose dormancy on a seed. In general, seed coat hardness is reflected in its lignin content, with hardness increasing with lignin content (Werker 1997). Interestingly, after retrieval of O. acanthium cypselas from the soil, those with a higher lignin content germinated more readily, possibly because they were more brittle and therefore more easily cracked, effectively mimicking scarification and allowing protrusion of the radicle and subsequent germination.

Cypselas retrieved from the seed bank after 3 years all showed a similar pattern and amount of soluble chemicals. This is in contrast to the negative correlation found between a large amount of water-soluble inhibitors (presumably soluble phenolics) and low germinability in fresh or stored O. acanthium cypselas (M. Qaderi, unpublished data). This result suggests that during the 3 years in the soil, chemicals that could have imposed dormancy were removed from cypselas by leaching or otherwise degraded.

It has been shown that surface wax content of seeds may impose dormancy (Rangaswamy & Nandakumar 1985). The higher total surface wax found in the non-germinated cypselas relative to the germinated ones (whether retrieved from the field as splits or from the incubator) suggests that the waxy layer is involved in the inhibition of germination, possibly by preventing inhibitor release through the cypsela coat (Werker 1997). This study shows a possible link between surface wax on the cypsela coat of O. acanthium and the maintenance of dormancy under both field and laboratory conditions.


Despite the complex emergence patterns, many more viable cypselas remained in the seed bank after 3 years than had germinated and emerged during that period (Figs 3 and 4). It is not therefore surprising that large populations of O. acanthium have arisen around London, Ontario, in areas where no flowering plants have been seen for 5 or more years. Further investigation has confirmed the presence of a seed bank in such areas (P. Cavers, personal observation). Years in which warm ripening conditions coincide with the largest part of the cypsela crop will lead to less dormancy in the population as a whole. This will create large numbers of seedlings and many splits in the first 2 to 3 years after cypsela maturation, but a smaller persistent seed bank in later years.

Intermittency in O. acanthium cypselas, which is a problem for those who wish to control this weed, can be caused, at least in part, by environmental conditions during ripening and after shedding (Qaderi & Cavers 2000b). When cypselas reach the soil, some germinate and some are incorporated into the seed bank. Even though many cypselas may originate from the same population and mature under the same conditions, each will have its own microenvironment in the seed bank. We have demonstrated that soil type and soil depth strongly affect this microenvironment and that microenvironments differ for cypselas entering the seed bank on different dates. Even in the fourth growing season (1999) there were large differences between collections, between depths of storage and between soils in the numbers of seedlings appearing. An obvious advantage for many species with lengthy periods of seed maturation is the great variation in dormancy of their seeds, leading to intermittent patterns of seedling emergence.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We thank the Natural Sciences and Engineering Research Council of Canada for financial support. The authors also thank Hashim Qaderi and Zakera Qaderi for their assistance with burying drinks bottles and cypselas in the soil, and Alexa Seal and Anisa Omar for helping to count emergence at the ESW research station. We appreciate comments on the manuscript from Dr Ken Thompson, Dr Lindsay Haddon and two anonymous referees.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Baker, H.G. (1989) Some aspects of the natural history of seed banks. Ecology of Soil Seed Banks (eds M.A.Leck, V.T.Parker & R.L.Simpson), pp. 921. Academic Press, San Diego, California.
  • Baskin, C.C. & Baskin, J.M. (1998) Seeds: Ecology, Biogeography, and Evolution of Dormancy and Germination. Academic Press, San Diego, California.
  • Baskin, J.M. & Baskin, C.C. (1995) Variation in the annual dormancy cycle in buried seeds of the weedy winter annual Viola arvensis. Weed Research, 35, 353362.
  • Bewley, J.D. & Black, M. (1994) Seeds: Physiology of Development and Germination, 2nd edn. Plenum Press, New York.
  • Blackshaw, R.E. (1992) Soil temperature, soil moisture, and seed burial depth effects on redstem filaree (Erodium cicutarium) emergence. Weed Science, 40, 204207.
  • Bradbeer, J.W. (1988) Seed Dormancy and Germination. Chapman & Hall, New York.
  • Cavers, P.B. & Benoit, D.L. (1989) Seed banks in arable land. Ecology of Soil Seed Banks (eds M.A.Leck, V.T.Parker & R.L.Simpson), pp. 309328. Academic Press, San Diego, California.
  • Cavers, P.B., Groves, R.H. & Kaye, P.E. (1995) Seed population dynamics of Onopordum over 1 year in southern New South Wales. Journal of Applied Ecology, 32, 425433.
  • Cavers, P.B. & Harper, J.L. (1966) Germination polymorphism in Rumex crispus and Rumex obtusifolius. Journal of Ecology, 54, 369382.
  • Cavers, P.B., Qaderi, M.M., Manku, R. & Downs, M.P. (2000) Intermittent germination: causes and ecological implications. Seed Biology: Advances and Applications (eds M.Black, K.J.Bradford & J.Vázquez-Ramos), pp. 363374. CAB International, Wallingford.
  • Doucet, C. & Cavers, P.B. (1996) A persistent seed bank of the bull thistle Cirsium vulgare. Canadian Journal of Botany, 74, 13861391.
  • Environment Canada (1996) Monthly Meteorological Summary. London, Ontario. Environment Canada, Ontario Region, Downsview, Ontario.
  • Environment Canada (1997) Monthly Meteorological Summary. London, Ontario. Environment Canada, Ontario Region, Downsview, Ontario.
  • Environment Canada (1998) Monthly Meteorological Summary. London, Ontario. Environment Canada, Ontario Region, Downsview, Ontario.
  • Environment Canada (1999) Monthly Meteorological Summary. London, Ontario. Environment Canada, Ontario Region, Downsview, Ontario.
  • Fenner, M. (1985) Seed Ecology. Chapman & Hall, London.
  • Fenner, M. (1991) The effects of the parent environment on seed germinability. Seed Science Research, 1, 7584.
  • Grundy, A.C. & Mead, A. (1998) Modeling the effects of seed depth on weed seedling emergence. Aspects of Applied Biology, 51, 7582.
  • Gutterman, Y. (2000) Seed dormancy as one of the survival strategies in annual plant species occurring in deserts. Dormancy in Plants: from Whole Plant Behaviour to Cellular Control (eds J.-D.Viémont & J.Crabbé), pp. 139159. CAB International, Wallingford.
  • Harper, J.L. (1977) Population Biology of Plants. Academic Press, London.
  • Lu, F. & Ralph, J. (1997) Derivatization followed by reductive cleavage (DFRC method), a new method for lignin analysis: protocol for analysis of DFRC monomers. Journal of Agricultural and Food Chemistry, 45, 25902592.
  • Malone, C.R. (1967) A rapid method for enumeration of viable seeds in soil. Weeds, 15, 381382.
  • McKenzie, J., Schaefer, R.L. & Farber, E. (1995) The Student Edition of Minitab for Windows, 2nd edn. Addison-Wesley, Reading, Massachusetts.
  • Milberg, P. & Andersson, L. (1998) Dormancy dynamics in buried weed seeds. Aspects of Applied Biology, 51, 153158.
  • Mohler, C.L. (1993) A model of the effects of tillage on emergence of weed seedlings. Ecological Applications, 3, 5373.
  • Parker, V.T., Simpson, R.L. & Leck, M.A. (1989) Pattern and process in the dynamics of seed banks. Ecology of Soil Seed Banks (eds M.A.Leck, V.T.Parker & R.L.Simpson), pp. 367384. Academic Press, San Diego, California.
  • Parsons, W.T. (1973) Noxious Weeds of Victoria. Inkata Press, Melbourne, Australia.
  • Priestley, D.A. (1986) Seed Aging: Implications for Seed Storage and Persistence in the Soil. Cornell University Press, Ithaca, New York.
  • Qaderi, M.M. (1998) Intraspecific variation in germination of Scotch thistle (Onopordum acanthium L.) cypselas. MSc thesis. Department of Plant Sciences, University of Western Ontario, London, Ontario.
  • Qaderi, M.M. & Cavers, P.B. (2000a) Variation in germination response among local populations of Scotch thistle, Onopordum acanthium L. Seed Science and Technology, 28, 881886.
  • Qaderi, M.M. & Cavers, P.B. (2000b) Variation in germination response within Scotch thistle, Onopordum acanthium L., populations matured under greenhouse and field conditions. Écoscience, 7, 5765.
  • Qaderi, M.M. & Cavers, P.B. (2002) Interpopulation and interyear variation in germination in Scotch thistle, Onopordum acanthium L., grown in a common garden: genetics vs environment. Plant Ecology, 162, 18 .
  • Rangaswamy, N.S. & Nandakumar, L. (1985) Correlative studies on seed coat structure, chemical composition and impermeability in the legume Rhynchosia minia. Botanical Gazette, 146, 501509.
  • Redmann, R.E. & Qi, M.Q. (1992) Impacts of seeding depth on emergence and seedling structure in eight perennial grasses. Canadian Journal of Botany, 70, 133139.
  • Roberts, H.A. (1981) Seed banks in soil. Advances in Applied Biology, 6, 155.
  • Roberts, H.A. & Chancellor, R.J. (1979) Periodicity of seedling emergence and achene survival in some species of Carduus, Cirsium and Onopordum. Journal of Applied Ecology, 16, 641647.
  • Roberts, H.A. & Feast, P.M. (1972) Fate of seeds of some annual weeds in different depths of cultivated and undisturbed soil. Weed Research, 12, 316324.
  • Salisbury, E.J. (1961) Weeds and Aliens. Collins, London.
  • SAS Institute Incorporation (1989) SAS User's Guide: Statistics. SAS Institute, Cary, North Carolina.
  • Smith, H.A., Johnson, W.S., Scott Shonkwiler, J. & Swanson, S.R. (1999) The implications of variable or constant expansion rates in invasive weed infestations. Weed Science, 47, 6266.
  • Sokal, R.R. & Rohlf, F.J. (1995) Biometry: the Principles and Practice of Statistics in Biology Research, 3rd edn. W.H. Freeman, New York.
  • Stabell, E., Upadhyaya, M.K. & Ellis, B.E. (1996) Development of seed coat-imposed dormancy during seed maturation in Cynoglossum officinale. Physiologia Plantarum, 97, 2834.
  • Taylorson, R.B. (1970) Changes in dormancy and viability of weed seeds in soils. Weed Science, 18, 265269.
  • Toole, E.H. & Brown, E. (1946) Final results of the Duvel buried seed experiment. Journal of Agricultural Research, 72, 201210.
  • Werker, E. (1997) Seed Anatomy. Gebrüder Borntraeger, Berlin, Stuttgart.
  • Williams, R.D. & Hoagland, R.E. (1982) The effects of naturally occurring phenolic compounds on seed germination. Weed Science, 30, 206212.
  • Zar, J.H. (1999) Biostatistical Analysis, 4th edn. Prentice Hall, Upper-Saddle River, New Jersey.