• hairy galinsoga;
  • gallant soldier;
  • burial depth;
  • emergence;
  • photoblasticity;
  • primary dormancy;
  • seed longevity


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

Galinsoga quadriradiata (hairy galinsoga) and Galinsoga parviflora (smallflower galinsoga, gallant soldier) are very troublesome weeds in many vegetable row crops in Europe. To optimise management strategies for Galinsoga spp. control, an in-depth study of germination biology was performed. Germination experiments were conducted to evaluate the impact of light and alternating temperatures on germination of a large set of Galinsoga populations. Seedling emergence was investigated by burying seeds at different depths in a sandy and sandy loam soil. Dormancy of fresh seeds harvested in autumn was evaluated by studying germination response in light at 25/20°C with and without nitrate addition. Seed longevity was investigated in an accelerated ageing experiment by exposing seeds to 45°C and 100% relative humidity. Galinsoga spp. seeds required light for germination; light dependency varied among populations. Seedling emergence decreased drastically with increasing burial depth. Maximum depth of emergence varied between 4 and 10 mm depending on soil type and population. In a sandy soil, emergence percentages were higher and seedlings were able to emerge from greater depths than in a sandy loam soil. Freshly produced G. parviflora seeds, harvested in autumn, showed a varying but high degree of primary dormancy and were less persistent than G. quadriradiata seeds that lack primary dormancy. Lack of primary dormancy of freshly harvested G. quadriradiata seeds and light dependency for germination may be used to optimise and develop Galinsoga management strategies.


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

The summer annuals Galinsoga quadriradiata Ruiz & Pavon (hairy galinsoga) and Galinsoga parviflora Cav. (smallflower galinsoga, gallant soldier) are considered to be native to South America, most likely Peru (Damalas, 2008) and are now found in many parts of the temperate and subtropical regions of the World. After being introduced in Europe, Galinsoga spp. were for a long time regarded as not problematic (Riemens & van der Weide, 2008). In Belgium, G. parviflora and G. quadriradiata have gradually spread since their first registered introduction in 1885 and 1929 (Van Landuyt et al., 2006) respectively. They are now major weeds on sandy and sandy loam soils in low growing, mainly vegetable, row crops with a slow ground cover, such as carrots (Daucus carota L.), onion (Allium cepa L.), black salsify (Scorzonera hispanica L.), Brussels endives (Cichorium intybus L. var. foliosum Hegi) and salad crops, such as lettuce (Lactuca sativa L.) and salad endive (Cichorium endivia L.) (Ivany & Sweet, 1973; Riemens & van der Weide, 2008). Both species are morphologically difficult to distinguish. The identification requires a careful study of the paleae at the base of the disc florets, the pappi of disc florets and stem hairiness with the aid of a hand lens. Galinsoga quadriradiata has untoothed paleae and some awned pappi and dense strongly divergent white hairs, whereas G. parviflora has three-toothed paleae, unawned pappi and is less hairy. Determination is further complicated by the occurrence of interspecific hybrids, resulting in phenotypical intermediates (Van Der Meijden, 2005).

The available pre-emergence and post-emergence herbicide portfolio for control of Galinsoga spp. in many minor crops is very limited (Hoek & van der Weide, 2007; Riemens & van der Weide, 2008) and continuously becomes smaller, owing to more stringent biocide EU regulations. Hand weeding and mechanical weeding of Galinsoga spp. is compromised, due to the presence of a dense and shallow root system with multiple secondary and adventitious roots (Canne, 1977). Moreover, both species are characterised by a number of biological characteristics that make them very prolific: the production of a great number of seeds in a wide range of environmental circumstances (Usami, 1976; Rai & Tripathi, 1983), the lack of seed dormancy of freshly produced seeds (Ivany, 1975; Warwick & Sweet, 1983; Martínez-Ghersa et al., 2000), a short life cycle with rapid early growth and development (Rai & Tripathi, 1983; Warwick & Sweet, 1983), early flowering (Pladeck, 1933; Warwick & Sweet, 1983), enabling several generations per growing season (Riemens & van der Weide, 2008), easy vegetative reproduction by adventitious root formation of clipped stems under moist conditions (Ivany, 1975; Warwick & Sweet, 1983; Damalas, 2008) and the ability to set seeds on uprooted or clipped flowering plants (van Poeteren, 1935). Only intensive weed control with several passes across fields is able to control these weeds.

Both Galinsoga species differ greatly in distribution and abundance, despite their morphological resemblance and relatedness. Despite being introduced 44 years later than G. parviflora, G. quadriradiata gradually became the most widespread and abundant Galinsoga species in Belgium (Van Landuyt et al., 2006). We hypothesised that one of the possible reasons for this differential pattern might be different seed biology. Likewise, if both species do differ in seed biology, the success of management strategies aiming to deplete seedbanks will be species-dependent. Hence, to explain the differential success of Galinsoga species in Belgium and to optimise preventive and curative integrated weed management strategies for Galinsoga control, an in-depth study of seed biology, particularly on seed germination, seed dormancy and seed longevity, is required. In particular, the following research questions were addressed: (i) Do G. quadriradiata and G. parviflora differ in their germination response to alternating temperatures and light conditions? (ii) What is the optimal and maximum soil depth for emergence of buried G. parviflora and G. quadriradiata seeds? (iii) Do freshly produced seeds of G. quadriradiata and G. parviflora exhibit primary dormancy? and (iv) Do G. quadriradiata and G. parviflora differ in seed longevity?

Materials and methods

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

Galinsoga spp. populations

As an interspecific difference in germination, dormancy and persistence of seeds may depend on the genetic constitution of the tested material, experiments were performed on a large set of Galinsoga spp. populations. Experiments were conducted in the summer and autumn of 2011 with seeds from local populations and from two reference populations, G. parviflora ‘Herbiseed1’ and G. quadriradiata ‘Herbiseed3’ purchased from the seed company Herbiseed (Twyford, UK) in 2010 (Table 1). Galinsoga parviflora ‘Herbiseed2’ seedlot was harvested in 2011 on plants grown in Melle from G. parviflora ‘Herbiseed1’ seeds to have freshly harvested seeds available at any time. These reference populations are international standards widely used for interpopulation comparison. Seeds of 15 Flemish populations were collected in summers of 2010 and 2011 from well-fertilised sandy fields that were located at least 4 km apart. Selected fields contained only one Galinsoga species; hence, we were sure not to work with interspecific hybrids. Fully mature seeds were collected randomly on at least 60 plants scattered over the whole field and pooled. Seeds were carefully removed from seed heads at full maturity with pappi of the seeds fully expanded. Seed populations were named after the nearby town where they were collected (Table 1). Seedlots used in Experiments 1 and 2 were all harvested in September 2010, except for G. parviflora ‘Herbiseed2’ and ‘Haasdonk’, and for G. quadriradiata ‘Haasdonk’ in Experiment 1 which were harvested in July 2011. Seeds harvested in 2010 were kept in hermetically sealed plastic boxes at 5°C before being used in the experiments. Only freshly harvested seeds (September–October 2011) were used in Experiments 3 and 4 (Table 1).

Table 1. Tested populations of Galinsoga parviflora and G. quadriradiata with harvest dates of the seeds
SpeciesTested populationHarvest date of the seedsTested in experimentsa
1 (08–09/2011)2 (08/2011)3 (10/2011)4 (10/2011)
  1. a

    Experimental period is given in brackets.

  2. b

    Standard reference population.

  3. c

    Harvested in 2011 on plants grown in Melle from G. parviflora ‘Herbiseed1’ seeds.

G. parviflora Destelbergen109/2010xx  
Destelbergen210/2011  xx
Haasdonk07/2011x x 
Herbiseed1 (UK)b09/2010xx  
Herbiseed2c07/2011x x 
Puurs09/2011  xx
Sint-Niklaas10/2011  xx
G. quadriradiata Beitem09/2010xx  
Gent09/2011  xx
Melle210/2011  xx
Nieuwkerken10/2011  x 
Puurs09/2011  x 
Sint-Niklaas10/2011  xx

Germination experiments (Exp. 1)

Germination of seedlots was evaluated under different alternating temperature and light regimes in a series of ‘top of paper’ (TP) germination tests. Tested seedlots comprised four G. quadriradiata (one reference and three local populations) and six G. parviflora populations (two references and four local populations) (Table 1). All seedlots were exposed to two light levels (24 h darkness, 16 h light/8 h darkness) and four alternating day/night temperature regimes 11/6°C, 15/10°C, 25/20°C and 35/30°C for 21 days. The alternating temperature regimes represent some seasonal day/night fluctuations in Belgium. The photoperiod represents the average day length from May to August in Belgium. Constant temperature regimes were not evaluated; indeed, Andersen (1968) found lower germination at constant temperatures than at alternating temperatures.

Four separate subexperiments, one for each temperature regime, were established to investigate the response of alternating temperatures on germination in darkness and alternating light. Each subexperiment was designed as a split plot, with the two light levels as the vertical treatments and the 10 Galinsoga populations as the horizontal treatments, in four replicates of 100 seeds each. Each replicate consisted of two filter papers (Rotilabo Type 112A; Carl Roth, Karlsruhe, Germany), loaded with 50 seeds. Filter papers were put on the germination plates of a Jacobsen germination table and moistened by paper wicks (Schleicher-Schuell, Dassel, Germany) extending into the underlying water bath. Filter papers were covered with a bell jar provided with a hole at the top to prevent drying out. The temperatures were conditioned directly by heating/cooling the water that circulated inside the hollow germination plates. Dark germination was achieved by covering the plots with light-tight black boxes that were removed for a short period of time (15 s on average) to register the number of germinated seeds. Hence, absolute darkness was not obtained. Germinated seeds were counted and removed every 5 days during a period of 3 weeks. Seeds were considered germinated when their seedlings produced healthy cotyledons.

Burial pot experiment (Exp. 2)

A seed burial pot experiment was conducted to investigate the relation between seed burial depth and seedling emergence. The trial was conducted in a growth chamber with a 16-h day–8-h night photoperiod and a light intensity of 100 μmol m−2 s−1 at the soil level provided by Gro-lux cool white fluorescent lamps (F58W/GRO-T8; Sylvania, Erlangen, Germany) with a red:far red (R:FR) ratio of 98. Daily daytime temperature increased from 15 to 25°C at a rate of 2.5°C per hour, was kept at 25°C for 8 h and decreased from 25 to 15°C at a rate of 2.5°C per hour. The daily night-time temperature was kept at 15°C. Pots were irrigated by overhead sprinklers (2.1 mm day−1) as frequently as needed. Seedlots of two local G. quadriradiata populations and three G. parviflora populations (Table 1) were buried in both sandy and sandy loam soils, representative of the vegetable-growing regions in Belgium, at 0, 1, 2, 3, 4, 5, 6, 8 and 10 mm below the surface. In a preliminary 50-day pot experiment, conducted in natural sunlight, no seedlings emerged from soil layers deeper than 8 mm. Two separate subexperiments, one for each soil type, were established. The experimental design was a completely randomised block design with all factorial combinations of 5 Galinsoga populations and nine burial depths in four replicates of 50 seeds each. The pots (7 cm high × 9.5 cm in diameter, with drainage outlets covered with strips of nylon mesh) were filled with steamed and sifted (through a 2 mm sieve) sandy loam (subexperiment 1a) containing 2.6% organic matter, 8.6% clay (<2 μm), 51.6% silt (2–50 μm) and 39.9% sand (>50 μm) with a pH-KCl of 5.7), and with steamed and sifted (through a 2 mm sieve) humic sand (subexperiment 1b) containing 4.1% organic matter, 0.9% clay, 4.1% silt and 95.0% sand with a pH-KCl of 5.4. Soils were poured into each pot up to the lower mark. Seeds were then placed on the substrate surface, and the pots were filled up to the upper mark with additional soil. Both marks confined the burial depth. Before and after seed placement, the uppermost layer of soil was levelled and then pressured with the same force to standardise burial depths. Emerged, live seedlings (designated as emergence percentage, i.e. percentage of buried seeds that developed into emerged seedlings with cotyledons visible at the soil surface) were counted and clipped every 5 days for 3 weeks.

Dormancy-breaking experiment (Experiment 3)

To investigate the effect of potassium nitrate on dormancy breaking and germination, an experiment was conducted using freshly harvested (7-day-old seeds) seed populations of five G. quadriradiata and five G. parviflora populations (Table 1). At the beginning of the TP germination test, all seedlots were exposed to two potassium nitrate levels: 0% and 0.2% KNO3 by saturating the filter papers, following the guidelines of the International Seed Testing Association (Hampton & TeKrony, 1995). Water was used for moistening thereafter. Seeds were germinated on the germination table under an alternating day/night temperature of 25/20°C and a 16 h light/8 h dark regime. This temperature regime gave the maximum germination in previous germination experiments. Galinsoga populations and potassium nitrate levels were arranged in a completely randomised design with four replicates of 100 seeds each. Counting of germinated seeds followed the protocol described in Experiment 1. Prior to the start of the experiment, viability of seeds was checked by cutting seeds (100/seedlot) lengthwise followed by microscopic examination of the endosperm (cutting test). Seeds were scored as viable if firm and white endosperm was present. Shrivelled, dark brown or mushy seeds were scored as unviable. The percentage of dead seeds was below 1%, irrespective of seed population.

Accelerated ageing experiment (Experiment 4)

P-indices calculated in an accelerated ageing experiment are considered good predictors of seed vigour and persistence (Hampton & TeKrony, 1995; Long et al., 2008). P50 and P90 persistence indices reflect incubation periods causing 50% and 90% reduction in viability. Seeds of three G. parviflora and three G. quadriradiata populations (Table 1) were placed in an incubator (Model 1535; Shel Lab, Sheldon Manufacturing, Cornelius, USA) operating at 100% relative humidity and 45°C, following the guidelines of the International Seed Testing Association (Hampton & TeKrony, 1995). Six incubation periods were tested: 0, 1, 2, 3, 4, 8 and 16 days. The seeds were washed after incubation. Four replicates per incubation time and per population were put in completely randomised blocks on the germination table and exposed to 21 days of alternating day/night temperature (25/20°C) under a 16 h light/8 h dark regime. Counting of germinated seeds followed the protocol described in Experiment 1. The validity of the test was confirmed after determination of seed water content before and after each ageing period. Viability of non-germinated seeds was checked by the cutting test mentioned in Experiment 3.

Data analysis

All germination data were expressed as percentages. All treatment combinations were treated as a split-plot design (Experiment 1) or randomised complete block design (Experiments 2 and 3) with four replicates using anova in R (R Development Core Team, 2010). Treatments comprised all combinations of ten Galinsoga populations with four temperature regimes and two light conditions (Experiment 1), five Galinsoga populations with eight burial depths and two soil substrates (Experiment 2), and ten Galinsoga populations and two nitrate levels (Experiment 3). Due to significant interactions between experimental factors (< 0.05), the data were split and compared within populations. To determine the significant differences between group means, the Tukey HSD test (for normally distributed data) or the Bonferroni test (for non-normally distributed data) was used.

The accelerated ageing experiment (Experiment 4) was treated as a completely randomised design with four replicates and two experimental factors: incubation period and population. Persistence indices P50 and P90 were calculated using the three-parameter Weibull model (Knezevic et al., 2007):

  • display math(1)

where Y is the viability of the seeds, x the incubation period (days), d is the upper limit (i.e. the initial viability) and b is the slope of the curve around the P50. Germination data were Box–Cox transformed to obtain variance homogeneity. Regression analysis of the accelerated ageing experiment was performed using the drc package in R (Ritz & Streibig, 2005; R Development Core Team, 2010).


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

Germination response to light and alternating temperatures (Experiment 1)

Both species had the greatest percentage of germination in the 16 h light photoperiod, irrespective of population or alternating temperature regime (Table 2). Germination was poor in darkness, irrespective of temperature regime (Table 3). In light, seeds of G. quadriradiata populations germinated best under the alternating temperature regime of 25/20°C with significantly higher germination percentages as compared with lower alternating temperatures 11/6°C and 15/10°C, except for ‘Beitem’ (Table 2), which germinated equally well under alternating temperatures of 15/10, 25/20 and 35/30°C. Contrary to G. quadriradiata populations, germination response of G. parviflora seeds to alternating temperatures in light was dependent on population.

Table 2. Seed germination (%) in light (16 h photoperiod) of Galinsoga parviflora and G. quadriradiata populations exposed to four alternating day/night temperature regimes (Experiment 1)
Species: populationAlternating temperature regime
  1. Mean values within a row for a given Galinsoga population followed by the same letter are not significantly different at P = 0.05 according to the Tukey HSD test (in case of homoscedasticity) or the Bonferroni test (in case of heteroscedasticity).

G. parviflora:
Destelbergen19.3 ± 1.80 c27.8 ± 2.25 ab37.5 ± 3.23 a20.8 ± 0.85 b
Haasdonk3.3 ± 0.75 b0.8 ± 0.25 c2.5 ± 0.65 c29.0 ± 1.29 a
Herbiseed113.0 ± 3.54 c68.8 ± 2.84 b88.0 ± 2.04 a65.5 ± 1.85 b
Herbiseed21.5 ± 0.65 c2.3 ± 1.11 bc13.8 ± 1.38 b64.3 ± 2.46 a
Nieuwkerken14.3 ± 4.31 c46.0 ± 4.02 b67.5 ± 5.69 a43.5 ± 5.91 b
Ursel8.3 ± 1.97 c12.8 ± 0.85 bc14.5 ± 1.55 b32.5 ± 2.18 a
G. quadriradiata:
Beitem62.5 ± 4.87 b86.0 ± 2.48 a93.5 ± 3.86 a81.0 ± 2.80 a
Haasdonk13.8 ± 2.53 b13.0 ± 8.04 b97.0 ± 1.22 a89.5 ± 4.37 a
Herbiseed314.8 ± 1.25 c68.5 ± 4.52 b83.0 ± 6.04 a69.3 ± 1.44 ab
Melle117.3 ± 2.87 c45.0 ± 0.91 b64.0 ± 3.89 a45.3 ± 2.10 b
Table 3. Seed germination (%) in darkness of Galinsoga parviflora and G. quadriradiata populations exposed to four alternating day/night temperature regimes (Experiment 1)
Species: populationAlternating temperature regime
  1. Mean values within a row for a given Galinsoga population followed by the same letter are not significantly different according to the Tukey HSD test (in case of homoscedasticity) or the Bonferroni test (in case of heteroscedasticity). No significant differences between figures with the same letter; comparison within rows only.

  2. Mean values are followed by *(< 0.05), **(< 0.01) or ***(< 0.001) in case germination in darkness was significantly different from germination in light, or by NS when differences were insignificant.

G. parviflora:
Destelbergen10.8 ± 0.48c NS17.5 ± 3.20a*7.0 ± 1.08b***1.8 ± 0.48bc***
Haasdonk0.0 ± 0.00a*0.0 ± 0.00a NS0.8 ± 0.25a NS0.0 ± 0.00a***
Herbiseed10.0 ± 0.00a**7.0 ± 2.12a***2.0 ± 0.00a***0.8 ± 0.48a***
Herbiseed20.0 ± 0.00a NS0.5 ± 0.50a NS1.3 ± 0.63a***0.3 ± 0.25a***
Nieuwkerken5.0 ± 1.87a NS23.3 ± 3.33a*12.0 ± 2.97a***4.3 ± 1.49a***
Ursel1.0 ± 0.71a*1.0 ± 0.71a***1.5 ± 0.50a***1.3 ± 0.48a***
G. quadriradiata:
Beitem37.5 ± 4.87b***87.0 ± 1.91a NS51 ± 3.70b***25.5 ± 1.55b***
Haasdonk0.3 ± 0.25c NS0.8 ± 0.48c NS21.3 ± 4.63a***7.0 ± 2.92ab***
Herbiseed30.0 ± 0.00a*10.0 ± 1.83a***7.5 ± 0.65a***0.8 ± 0.48a***
Melle11.8 ± 0.85b**17.5 ± 3.43a***24.3 ± 1.31a***1.8 ± 0.48b***

Within alternating temperature regimes, germination of Galinsoga spp. populations was significantly lower in darkness than in the light (Table 3), except for G. quadriradiata ‘Beitem’ at 15/10°C and provided that germination in light exceeded 15%. In darkness, germination percentages of Galinsoga spp. populations were lower than 25% except for G. quadriradiata ‘Beitem’, which germinated for 87% at 15/10°C (Table 3). In darkness, alternating temperatures did not affect germination of G. quadriradiata ‘Herbiseed3’ and all G. parviflora populations, except for population ‘Destelbergen’.

Response of seedling emergence to burial (Experiment 2)

Seedling emergence decreased exponentially with increasing burial depth, irrespective of soil type used (Table 4), with highest emergence for unburied seeds.

Table 4. Emergence (%) of three Galinsoga parviflora and two G. quadriradiata populations as affected by depth of sowing (mm) in sandy loam (SL) and sandy (S) soil (Exp. 2)
Burial depth (mm) G. parviflora G. quadriradiata
  1. a

    SL = sandy loam soil.

  2. b

    S = sandy soil.

  3. c

    Significant difference between soil substrates according to Tukey HSD test (in case of homoscedasticity) or Bonferroni test (in case of heteroscedasticity): NS, non-significant; *< 0.05; **< 0.01; ***< 0.001.

042.5 ± 3.5931.0 ± 3.42NS96.0 ± 2.4589.5 ± 3.20NS63.5 ± 5.5648.5 ± 3.59NS76.0 ± 15.3877.5 ± 5.85NS56.0 ± 6.0050.5 ± 3.30NS
127.5 ± 1.5020.0 ± 4.69NS51.3 ± 8.5079.0 ± 1.73*10.0 ± 3.0045.3 ± 8.33***74.5 ± 13.5971.5 ± 9.54NS10.5 ± 7.0938.5 ± 5.44***
223.0 ± 5.4515.5 ± 4.65NS38.0 ± 8.1856.5 ± 4.92NS10.7 ± 3.789.0 ± 3.11NS64.7 ± 16.7763.5 ± 1.71NS12.5 ± 3.8616.5 ± 2.50NS
310.5 ± 6.1812.0 ± 4.69NS22.0 ± 9.8051.0 ± 7.33**2.0 ± 0.815.3 ± 3.06NS44.0 ± 12.2966.7 ± 6.66NS3.0 ± 3.005.5 ± 3.10NS
40.7 ± 0.582.5 ± 1.50NS0.5 ± 0.507.3 ± 3.06NS2.0 ± 1.411.0 ± 1.00NS16.7 ± 8.3870.0 ± 4.24**1.0 ± 1.004.0 ± 3.37NS
51.0 ± 1.002.5 ± 1.50NS0.0 ± 0.002.0 ± 1.73NS0.0 ± 0.000.0 ± 0.00NS10.0 ± 4.5826.7 ± 2.31NS0.5 ± 0.504.0 ± 2.31NS
61.5 ± 1.501.5 ± 1.50NS0.5 ± 0.501.5 ± 0.50NS0.0 ± 0.000.5 ± 0.50NS3.5 ± 0.968.0 ± 3.61NS0.5 ± 0.500.5 ± 0.50NS
80.0 ± 0.001.5 ± 1.50NS0.0 ± 0.000.0 ± 0.00NS1.0 ± 1.000.0 ± 0.00NS0.5 ± 0.506.5 ± 4.27NS1.0 ± 0.580.0 ± 0.00NS
100.0 ± 0.000.0 ± 0.00NS0.0 ± 0.000.5 ± 0.50NS0.5 ± 0.500.0 ± 0.00NS0.5 ± 0.501.5 ± 0.96NS0.0 ± 0.000.0 ± 0.00NS

The influence of burial depth was extremely dependent on population. Emergence of G. parviflora populations ‘Nieuwkerken’ and ‘Herbiseed1’, and G. quadriradiata population ‘Melle1’, dropped drastically as the burial depth was increased from 0 to 2 mm in sandy soil and from 0–1 mm in sandy loam soil. For all other populations, this significant drop in seedling emergence occurred at deeper burial depths: for G. parviflora ‘Destelbergen1’ at 3 and 2 mm in sandy soil and sandy loam soil, respectively, whereas for G. quadriradiata ‘Beitem’ at 5 and 4 mm in sandy soil and sandy loam soil respectively. So, compared with sandy soil, emergence in sandy loam dropped at more superficial burial depths.

Maximum depth of emergence (defined as burial depth at which emergence is higher than 1%) varied among populations. Seedling emergence of Galinsoga parviflora ‘Nieuwkerken’, ‘Herbiseed1’ and ‘Destelbergen1’ occurred from a maximum depth of 3, 6 and 8 mm, respectively, in sandy soil and to 4, 3 and 6 mm, respectively, in sandy loam. Galinsoga quadriradiata ‘Melle1’ and ‘Beitem’ had maximum emergence depths of 5 and >10 mm, respectively, in sandy soil, and 3 and 6 mm, respectively, in sandy loam. With regard to soil types, significant differences in emergence within burial depths were found among populations, except for G. parviflora ‘Destelbergen1’. For G. parviflora ‘Herbiseed1’ and ‘Nieuwkerken’, and for G. quadriradiata ‘Melle1’ buried at a depth of 1 mm, higher emergence was found in sandy soil than in sandy loam soil. Likewise, G. quadriradiata ‘Beitem’ seeds buried at 4 mm emerged better in sandy soil than in sandy loam soil.

Degree of dormancy of freshly harvested seeds (Experiment 3)

The addition of potassium nitrate to the germination substrate did not significantly affect germination response in light at 25/20°C, irrespective of population, except for G. parviflora ‘Haasdonk’, which showed higher germination when seeds were exposed to potassium nitrate. Germination response of freshly harvested seeds to 25/20°C was significantly lower for G. parviflora populations than for G. quadriradiata populations, irrespective of potassium nitrate level (Fig. 1). Within G. parviflora, ‘Haasdonk’, ‘Herbiseed2’ and ‘Puurs’ germinated significantly better than ‘Destelbergen’ and ‘Sint-Niklaas’, irrespective of potassium nitrate level. Contrary to G. parviflora populations, no differences in germination response were found among G. quadriradiata populations.


Figure 1. Germination percentage (±SE) of freshly harvested seeds of five Galinsoga parviflora and five G. quadriradiata populations exposed to two potassium nitrate levels (without nitrate; with nitrate = 0.2% KNO3) under an alternating day/night temperature of 25/20°C and a 16 h light/8 h dark regime (Exp. 3).

Download figure to PowerPoint

Seed persistence of freshly harvested seeds (Experiment 4)

Initial germinability exceeded 85% for all populations except for G. parviflora ‘Destelbergen2’ (Table 5). Germinability of seed populations decreased with increasing incubation time at 45°C and 100% relative humidity. All seeds of G. parviflora and G. quadriradiata were dead after an incubation time of 4 and 8 days respectively (Table 5). Populations of G. quadriradiata were more persistent than populations of G. parviflora, as indicated by their significantly higher P50 and P90 values (Table 6), except for G. quadriradiata ‘Sint-Niklaas’. Within Galinsoga species, populations significantly differed in P90 value.

Table 5. Germination percentages with standard errors for three Galinsoga parviflora and three G. quadriradiata populations exposed to 45°C and 100% RH for 0, 1, 2, 4, 8 and 16 days in an incubator (Experiment 4)
Incubation period (days) G. quadriradiata G. parviflora
095.0 ± 0.5897.5 ± 1.5096.0 ± 2.3166.0 ± 2.9486.0 ± 0.8297.5 ± 0.96
188.5 ± 2.8797.0 ± 1.9187.0 ± 4.5140.5 ± 7.9373.5 ± 3.3072.5 ± 2.63
284.5 ± 3.8690.5 ± 2.5070.0 ± 5.290.0 ± 0.0023.0 ± 3.1148.5 ± 1.26
47.5 ± 1.260.5 ± 0.500.0 ± 0.000.0 ± 0.000.0 ± 0.000.0 ± 0.00
80.0 ± 0.000.0 ± 0.000.0 ± 0.000.0 ± 0.000.0 ± 0.000.0 ± 0.00
160.0 ± 0.000.0 ± 0.000.0 ± 0.000.0 ± 0.000.0 ± 0.000.0 ± 0.00
Table 6. Persistence-indices P50 and P90 (days) with standard errors for three Galinsoga parviflora and three G. quadriradiata populations kept at 45°C and 100% relative humidity (Exp. 4)
SpeciesPopulationP50 (days)P90 (days)
  1. No significant differences between figures with the same letter (based on computed selectivity indices and corresponding P-values); comparison within columns only.

G. quadriradiata Ekkergem3.1 ± 0.28 a3.9 ± 0.06 b
Melle22.8 ± 0.26 a3.3 ± 0.17 c
Sint-Niklaas2.2 ± 0.05 ab2.5 ± 0.10 a
G. parviflora Destelbergen21.0 ± 0.63 c1.0 ± 0.63 d
Puurs1.7 ± 0.07 b2.3 ± 0.06 a
Sint-Niklaas2.1 ± 0.07 b2.6 ± 0.11 a


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

Our experiments indicated that germination of Galinsoga spp. seeds occurred over a wide range of alternating temperatures (between 11/6 and 35/30°C), in line with the results of Damalas (2008) who found that Galinsoga spp. seeds can germinate within a range of 10–35°C. According to Ivany and Sweet (1973), Jursik et al. (2003) and Damalas (2008), minimum temperature for germination is 10°C. In our experiments, optimal alternating temperature regime for germination of Galinsoga spp. populations in light was 25/20°C. According to Shontz and Shontz (1972), 24/18°C was the optimal alternating temperature regime for germination of G. quadriradiata. However, G. parviflora ‘Ursel’, ‘Haasdonk’ and Herbiseed2' showed highest germination at 35/30°C. Their seeds may have shown a high degree of (primary or secondary) dormancy, which could only be partly overcome by exposing them to higher temperatures. Lowest germination in the light was found at alternating temperatures of 11/6°C. Germination in darkness was less dependent on alternating temperatures. Differences in germination among Galinsoga spp. populations at suboptimal germination temperatures may point to differential degrees of secondary dormancy among populations. These differences were most pronounced for G. quadriradiata ‘Beitem’ exhibiting high germination even at low temperatures. This population probably exhibited a low degree of secondary dormancy. According to Vleeshouwers et al. (1995), the level of dormancy determines the conditions needed for germination.

Galinsoga quadriradiata and G. parviflora seeds had a strong light dependency for germination: they germinated poorly under darkness, irrespective of alternating temperatures. This more or less conforms to results from Warwick and Sweet (1983) and Karlsson et al. (2008), who found no germination of Galinsoga spp. seeds in complete darkness. The low germination levels we found in darkness may be partly attributed to short-duration exposure to light every 5 days to allow the counting of germinated seeds. Compared with germination under absolute darkness, Andersson et al. (1997) found a significant increase in germination when seeds had been exposed for 5 s to a light intensity of 230 μmol m² s−1. Germination of G. quadriradiata ‘Beitem’ was clearly less dependent on light conditions than all other populations. ‘Beitem’ seeds exposed to alternating temperatures of 25/20°C and darkness for 21 days germinated up to 51%. Moreover, 10–27% of them, depending on soil type, emerged when buried at a soil depth of 5 mm and exposed to 25/20°C, whereas no or little emergence was found for all other populations buried at the same depth. These findings do not conform to those of Ivany and Sweet (1973), who found that emergence of G. quadriradiata was reduced from 76% to 1% when seeds were buried at a depth of 5 mm. The reaction of ‘Beitem’ was confirmed in a preliminary germination experiment, where six populations, including ‘Beitem’, were exposed to undisturbed absolute darkness for 21 days at 25/20°C; no germination was found, except for ‘Beitem’ of which 30% of the seeds were germinated.

Similar to findings of Warwick and Sweet (1983), Galinsoga seedlings primarily emerged from more superficial soil layers with highest emergence from unburied seeds. In general, emergence of Galinsoga spp. populations reduced with increasing burial depth, with no or little seedling emergence from depths deeper than 4 mm. This strong reduction indicates that both species require a minimum amount of light for germination and/or an appropriate light quality. Benvenuti and Macchia (1995) and Riemens et al. (2004) reported that light penetration fell below 0.01% at a depth of no more than 4 mm and that with increasing soil depth, light permeability was proportional to wavelength, leading to a progressive decline in R:FR ratio. It is well known that light with a high R:FR ratio enhances the formation of FR-absorbing phytochrome, which is required to break dormancy in many light-requiring species. However, the decline in emergence may also partly reflect species inability to produce a long seedling due to low seed energy reserves, as reported by Benvenuti et al. (2001). Unfortunately, we did not quantify the soil light penetration, nor did we check the presence of ungerminated seeds and unemerged seedlings (i.e. etiolated seedlings remaining below the soil surface). Our data were obtained under ideal environmental growth chamber conditions with regular overhead irrigation. A similar pattern is expected under natural field circumstances in moist soils with regular rainfall. Under dry conditions, optimal emergence depth probably will be less superficial. However, any inference regarding natural field circumstances should be carried out with care because of the 20-fold higher light intensity (about 2000 μmol m−2 s−1 for direct sunlight) and 90-fold lower R:FR ratio (approximately 1.1 for natural daylight) of direct sunlight as compared with the artificial illumination we used. The light intensities used in our study match light intensities on a shady day or under airy crop canopies. Seedling emergence and maximum emergence depth was dependent on soil type. In a sandy soil, emergence percentages were higher and seedlings were able to emerge from deeper depths than in sandy loam. This may be explained by the better light penetration in coarse grained soils (Tester & Morris, 1987). Rai and Tripathi (1983) also found the highest emergence of both Galinsoga species in soils with the highest sand fraction.

Huge interspecific differences in final germination percentages of freshly harvested seeds were found, contrary to findings of Ivany and Sweet (1973). In our experiments, newly formed G. quadriradiata seeds germinated immediately with germination percentages higher than 93%. These results conform with findings of Ivany (1975) and Reinhard et al. (2003), but contradict findings of Jursik et al. (2003) who found that G. quadriradiata seeds were primary dormant for 10 to 100 days, depending on time of seed ripening; seeds that ripened early in the season (July) showed longest dormancy duration. In our study, fresh seeds were harvested in autumn, which may explain the lack of primary dormancy. We cannot exclude that G. quadriradiata seeds, freshly harvested early in the season, exhibit a low degree of primary dormancy that may be relieved rapidly by high (germination) temperatures. This might explain why seeds of G. quadriradiata ‘Haasdonk’, harvested in summer 2011, germinated far less at 15/10°C (but not at higher temperatures) than G. quadriradiata seeds collected in autumn 2010 (see Table 2).

According to Ivany and Sweet (1973), Ivany (1975), Warwick and Sweet (1983) and Martínez-Ghersa et al. (2000), freshly harvested seeds of G. parviflora are non-dormant. However, in our study, freshly harvested seeds of G. parviflora germinated poorly in light with germination percentages between 10% and 80% and thus much less than in G. quadriradiata, depending on population. Seedlots ‘Haasdonk’ and ‘Herbiseed2’ that were harvested 2 months earlier than ‘Sint-Niklaas’, ‘Destelbergen2’ and ‘Puurs’ (see Table 1) showed two- to sevenfold higher germination levels than the freshly harvested ‘Sint-Niklaas’, ‘Destelbergen2’ and ‘Puurs’. Some seeds of ‘Haasdonk’ and ‘Herbiseed2’ might have gone through a process of after-ripening at the start of the experiment, underpinning the hypothesis of the existence of primary dormancy and a need for seed after-ripening to enable germination. Indeed, only 2 months before the start of Experiment 3, germination percentages of ‘Haasdonk’ and ‘Herbiseed2’ were 6- to 25-fold lower. Addition of potassium nitrate could not replace the requirement for after-ripening. This conforms to Riemens et al. (2004) who found no effect of nitrate addition on dormancy breaking.

Galinsoga quadriradiata seed populations were more persistent than G. parviflora seed populations when exposed to 45°C and 100% relative humidity, as shown by their higher P90 values. So, seed longevity in soil may be higher for G. quadriradiata than for G. parviflora. Indeed, as shown by Long et al. (2008), the P90 index is well correlated with seed viability of Galinsoga spp. seeds in soil. According to Roberts (1986), Hoek and van der Weide (2007) and Damalas (2008), seeds of Galinsoga buried in soil remain viable for at least 2 years. Espinoza-Garcia et al. (2003) reported that most G. parviflora seeds loose their viability within the first 10 months after burial.


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

Galinsoga spp. populations were able to germinate over a wide range of alternating temperatures (between 11/6 and 35/30°C). Seeds germinated at shallow depths, reflecting a high light dependence with large intraspecific differences in light requirement. The light dependency for germination may be used to optimise Galinsoga control. Burial of seeds by tillage operations may reduce field emergence, provided that Galinsoga seeds returned to the soil surface are less viable. Working depth of mechanical weeding machines should be as superficial as possible to minimise emergence of deeply buried seeds. In contrast with G. parviflora, freshly produced seeds of G. quadriradiata were able to germinate soon after shedding, owing to the lack of primary dormancy. Combined with a short life span between germination and flowering, multiple generations of G. quadriradiata can germinate, mature and produce seed within the same growing season. Through this prolific seed production, resulting in a large viable seedbank, reinfestation remains problematic for many years. Emergence may be reduced by depletion of soil seedbanks using stale seedbed preparations, by living mulches or by installing dead mulches, provided that mulch thickness is at least 5 mm. The longevity of seeds differs in both species, with G. parviflora showing smaller values. In line with this, depletion of seedbank of G. parviflora is expected to be faster, but it may be counteracted by the higher degree of dormancy. The lack of primary dormancy and high seed persistence may explain the higher distribution and abundance of G. quadriradiata over G. parviflora populations in Belgium, as reported by Van Landuyt et al. (2006). We conclude that inter- and intraspecific differences in seed germination, seed dormancy and seed longevity will determine the effectiveness of the suggested weed management strategies.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References
  • Andersen AN (1968) Germination and Establishment of Weeds for Experimental Purposes. W.F. Humphrey Press Inc., New York, USA.
  • Andersson L, Milberg P & Noronha A (1997) Germination response of weed seeds to light, light of short duration and darkness after stratification in soil. Swedish Journal of Agriculture 27, 113120.
  • Benvenuti S & Macchia M (1995) Effect of hypoxia on buried weed seed germination. Weed Research 35, 343351.
  • Benvenuti S, Macchia M & Miele S (2001) Quantitative analysis of emergence of seedlings from buried weed seeds with increasing soil depth. Weed Science 49, 528535.
  • Canne JM (1977) A revision of the genus Galinsoga (Compositae: Heliantheae). Rhodora 79, 319389.
  • Damalas CA (2008) Distribution, biology, and agricultural importance of Galinsoga parviflora (Asteraceae). Weed Biology and Management 8, 147153.
  • Espinoza-Garcia FJ, Vazquez-Bravo R & Martinez-Ramos M (2003) Survival, germinability and fungal colonization of dimorphic achenes of the annual weed Galinsoga parviflora buried in the soil. Weed Research 43, 269275.
  • Hampton JG & TeKrony DM (1995) Handbook of Vigour Test Methods. The International Seed Testing Association, Zürich, Switzerland.
  • Hoek J & van der Weide RY (2007) Bestrijding van Knopkruid, Ooievaarsbek en Reigersbek. Praktijkonderzoek Plant & Omgeving B.V., Wageningen, The Netherlands.
  • Ivany JA (1975) Today's weed: Galinsoga. Weeds Today 6, 22.
  • Ivany JA & Sweet RD (1973) Germination, growth, development and control of Galinsoga. Weed Science 21, 4145.
  • Jursik M, Soukup J, Venclova V & Holec J (2003) Seed dormancy and germination of Shaggy soldier (Galinsoga ciliata Blake.) and Common lambsquarter (Chenopodium album L.). Plant Soil Environment 49, 511518.
  • Karlsson LM, Tamado T & Milberg P (2008) Inter-species comparison of seed dormancy and germination of six annual Asteraceae weeds in an ecological context. Seed Science Research 18, 3545.
  • Knezevic SZ, Streibig JC & Rits C (2007) Utilizing R software package for dose-response studies: the concept and data analysis. Weed Technology 21, 840848.
  • Long RL, Panetta DF, Steadman KJ et al. (2008) Seed persistence in the field may be predicted by laboratory-controlled aging. Weed Science 56, 523528.
  • Martínez-Ghersa MA, Ghersa CM, Benech AR, MacDonough R & Sanchez RA (2000) Adaptive traits regulating dormancy and germination of invasive species. Plant Species Biology 15, 127137.
  • Pladeck M (1933) Report of the subcommittee on time of flowering and viability of weed species. Proceedings of Association of Official Seed Analysts 26, 249254.
  • van Poeteren N (1935) Het knopkruid, Galinsoga parviflora Cav. European Journal of Plant Pathology 41, 7476.
  • R Development Core Team (2010) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL
  • Rai JN & Tripathi R (1983) Population regulation of Galinsoga ciliata and G. parviflora. Weed Research 23, 151163.
  • Reinhard F, Herle M, Bastiansen F & Streit B (2003) Economic Impact of the Spread of Alien Species in Germany. Available at: (last accessed 18 October 2013).
  • Riemens MM & van der Weide RY (2008) Biology and control of Galinsoga parviflora, Overview of a literature survey. Plant Research International B.V., Wageningen.
  • Riemens MM, Scheepens PC & van der Weide RY (2004) Dormancy, Germination and Emergence of Weed Seeds, with Emphasis on the Influence of Light. Plant Research International B.V., Wageningen, The Netherlands.
  • Ritz C & Streibig JC (2005) Bioassay analysis using R. Journal of Statistical Software 12, 122.
  • Roberts HA (1986) Seed persistence in soil and seasonal emergence in plant species from different habitats. Journal of Applied Biology 23, 639656.
  • Shontz NN & Shontz JP (1972) Rapid evolution in populations of Galinsoga ciliata (Compositae) in western Massachusetts. American Midland Naturalist 88, 183199.
  • Tester M & Morris C (1987) The penetration of light through soil. Plant, Cell and Environment 10, 281286.
  • Usami Y (1976) Ecological studies on weeds in mulberry fields. 2. Auto-ecology of Galinsoga parviflora Cav. Weed Research (Japan) 21, 7680.
  • Van Der Meijden R (2005) Heukels' Flora of the Netherlands. Wolters-Noordhoff, Groningen, the Netherlands.
  • Van Landuyt W, Hoste I, Vanhecke L, Van Den Brempt P, Vercruysse W & De beer D (2006) Atlas van de Flora van Vlaanderen en het Brussels Gewest. Instituut voor Natuur- en Bosonderzoek, Nationale Plantentuin van België & Flo. Wer., Belgium.
  • Vleeshouwers LM, Bouwmeester HJ & Karssen CM (1995) Redefining seed dormancy: an attempt to integrate physiology and ecology. Journal of Ecology 83, 10311037.
  • Warwick SI & Sweet RD (1983) The Biology of Canadian Weeds. 58. Galinsoga parviflora and G. quadriradiata (= G. ciliata). Canadian Journal of Plant Science 63, 695709.