• emergence;
  • tropical;
  • shifting cultivation;
  • canopy shade;
  • fallow period;
  • invasive;
  • plant diversity


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Appendix 1: Species recorded in soil seedbank and corresponding surface vegetation with indication of abundance. Sites are four slash-and-burn rice fields in northern Laos. Field recordings were made in permanent quadrats from burning to harvest, 0 = no occurrence. The coppicing species are indicated: [r] = mainly resprouting sometimes seedling, [R] = only resprouting.

Crops in shifting cultivation fields often suffer from severe weed infestation when long fallow periods are replaced by short fallow periods. The soil seedbank as a source of weed infestation was studied in four fields that differed in their last fallow duration. The effect of burning was analysed by comparing adjacent pre-burn and post-burn samples (two sites). Surface vegetation was monitored from burning to harvest in the plots from which soil samples were taken to determine the fraction of the seedbank germinating (three sites). Seedbank size (1700–4000 seedlings m−2) varied depending on a single species, Mimosa diplotricha. Burning reduced emergence of most species, but stimulated emergence in others. Densities in the seedbank were not correlated with above-ground abundances in the field, except for some species. Most species emerging after 50 days from the soil samples (40% of seedlings) were absent from the field after 190 days. Whilst the data from this study are derived from only four fields, the weed problems after short-term fallowing appeared to be due to a larger fraction of the seedbank emerging, possibly due to shallow burial, and to a floristic shift towards adaption to burning, rather than the size of the seedbank per se.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Appendix 1: Species recorded in soil seedbank and corresponding surface vegetation with indication of abundance. Sites are four slash-and-burn rice fields in northern Laos. Field recordings were made in permanent quadrats from burning to harvest, 0 = no occurrence. The coppicing species are indicated: [r] = mainly resprouting sometimes seedling, [R] = only resprouting.

Slash-and-burn cultivation, commonly part of shifting cultivation, assures the subsistence of poor and rural populations, particularly in the tropics. Worldwide, approximately 190 million people practice shifting cultivation involving an area of about 1.5 million km2 (Plagge et al., 2008). Natural vegetation is cut and burned, and a single crop, rarely two, is planted, followed by a period in which the field reverts to natural regrowth, often secondary forest. Farmers apply no fertilisers and rely on fallow periods to restore soil fertility and suppress weeds. Harmful weeds are largely absent when long fallow periods and short cropping intervals are maintained, but under short fallows, weeds, more than declining soil fertility, tend to limit crop yields (Roder et al., 1997; Bech Bruun et al., 2006; Saito et al., 2006). Short fallows result from increased pressure on arable land. In Laos, this is exacerbated by two government policies (Lestrelin et al., 2005): resettlement of population near roads and land allocation; the latter provides but few parcels to farmers, thereby reducing fallow periods to 2–4 years (de Rouw et al., 2005). A succession of short fallows often leads to unmanageable weed invasion and fields being abandoned. There is no scientific evidence that seedbank size directly relates to fallow duration (Kellman, 1974; Rico-Gray & Garcia-Franco, 1992; Van Keer, 2003), and other factors, for instance, floristic composition and the seedbank response to weed control, should be considered in the relationship between seedbank and field infestation.

Soil seedbanks in shifting cultivation systems are of moderate size, 1000–5000 viable seeds m−2, compared with seed densities of over 10 000 seeds m−2 in tropical soils that are permanently cultivated with annual crops and densities below 300 seeds m−2 in uncultivated forest soil (Kellman, 1974; Rico-Gray & Garcia-Franco, 1992; Guevara et al., 2005). The fraction of weeds in the seedbank increases sharply with short rotations, by up to 100% in soil under permanent annual cropping (Garwood, 1989). Where slash-and-burn is the dominant land use, only a fraction of the land is cultivated in a given year, thus limiting the weed seed production in the locality. The seedbank is likely to include both weeds and forest species, because crops and fallow vegetation alternately occupy the site. The soil seedbank is both a source of weed infestation and of fallow restoration.

Burning of the natural vegetation destroys a proportion of seeds. The more biomass is burnt, the more likely seeds on the soil surface and in the top soil layer are killed by the fire. Although shifting cultivation is a no-till system, with burning leaving the soil ready to sow, some manual tillage may be performed to control weeds, associated with shorter fallow periods (Kunstadter & Chapman, 1978; Roder et al., 1997). In northern Laos, weed control is a combination of hand pulling the larger weeds, cutting resprouting stumps with a machete and spot weeding with a small curved hand hoe to cut the stems of seedlings above the soil surface. When seedlings are too numerous and too small to be cut, the same weeding tool is used to uproot weeds by scraping the soil surface (Dupin et al., 2009).

The objective of this study was to explore the importance of the seedbank as a source of weed infestation in traditional slash-and burn rice fields. For this purpose, we compared the emergence from soil samples with emergence in the rice field plots from which samples were taken. Positive and negative effects of shade, burning and hand weeding on emergence were measured, and the possible sources of dissimilarity in emergence of seedbank and field are discussed.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Appendix 1: Species recorded in soil seedbank and corresponding surface vegetation with indication of abundance. Sites are four slash-and-burn rice fields in northern Laos. Field recordings were made in permanent quadrats from burning to harvest, 0 = no occurrence. The coppicing species are indicated: [r] = mainly resprouting sometimes seedling, [R] = only resprouting.

Study area

The mean annual rainfall at Luang Prabang is 1403 mm, with approximately 91% falling between April and October. Total annual rainfall recorded at Ban Lak Sip, 10 km from Luang Prabang, during the period of study (2004) was 1383 mm. Northern Laos is predominantly mountainous. Alfisols are the most commonly cultivated soils; these are moderately fertile, 50 to over 200 cm deep clay soils with a loam or clay loam overlying top soil (MSEC, 1999). The natural vegetation is evergreen lowland forest on moist sites and semi-deciduous lowland forest often with bamboo on drier sites. Dipterocarp forest dominates the driest sites. Above 800 m, the appearance of Fagaceae indicates the transition from lowland to highland forest. In the study area, young fallow vegetation has replaced most of the original forest, but fallows remain species rich and rapidly acquire a forest structure. The majority of farmers in the study area are Khamou, traditional shifting cultivators who use the 250–1000 m elevation belt and the 20–55% slope range for cultivation, and who avoid dipterocarp forest.

Study sites

Sites were selected to be as uniform and representative as possible: elevation 300–600 m, soil type alfisol, soil depth >1 m, texture clay loam and original forest semi-deciduous. Only Khamou fields with rotations of fallow and upland rice were considered, thus excluding occasional crops of maize and Jobs' Tears (Coix lacryma-jobi L.). The four sites selected differed in cultivation frequency and fallow vegetation (Table 1).

Table 1. Characteristics of sample sites and experiments
  1. a

    Metallic cylinder 100 cm3, average over three samples.

Nearest villageHouay KhotNok PitBan Lak SipBan Lak Sip
Location field









Last fallow period (years)10650
Previous cultivation years1992, 19931983, 19971992, 1995, 19981987, 1991, 2000, 2003
Elevation above sea level (m)350640610540
Slope, mean of four plots (%)27586156
Bulk density (Mg m−3)a1.
Soil samples collected 2004
After slashing, before burn23 March24 March
After burn29 April7 April29 March25 March
Seedbank observations
Start nurseries30 April8 April31 March31 March
End nurseries31 August31 August31 August31 August
Field observations
Start with burning7 April29 March25 March
End with rice harvest10 October10 October10 October
Sites involved in experiments
Burning effectMeasuredMeasured
Sun-shade effectMeasuredMeasuredMeasuredMeasured
Seedbank vs. field emergenceMeasuredMeasuredMeasured
Scraping effectMeasured

Field S10 was covered by a 10-year old secondary forest forming a continuous canopy 15–20 m high and deep shade. Leuceana leucocephala (Lamk.) de Wit planted in 1993 accounted for about half of the canopy; the remaining cover was dominated by bamboo, Dendrocalamus strictus (Roxb.) Nees. Field S6 was covered by a 6-year-old secondary forest dominated by Bauhinia and Albizia trees, forming a continuous canopy at 6–12 m height and deep shade. Field S5 was covered by a 5-year-old secondary forest forming a continuous canopy at 5–10 m height and deep shade. Dominant trees of the upper layer were Callerya atropurpurea (Wall.) Schot and Sterculia lanceolata Cav. with Bauhinia and Albizia trees. Canopy trees were measured and identified as they lay on the ground after felling, shade conditions were judged from unfelled strips of the same forest vegetation. Field S0 was a field cultivated the previous year and covered by mixed vegetation 1.5 m high dominated by Chromolaena odorata (L.) R.M.King and H.Rob. overgrown with vines of Lepistemon binectariferum (Wall.) O. K. and Thunbergia grandiflora Roxb. with 3 m high tufts of Saccharum spontaneum L. Soil samples were collected at the end of the dry season, when seed dispersal peaks and germination have not yet begun, and seedbank density is at its annual maximum. Vegetation at all sites was slashed and burnt in the dry season of 2004 and cultivated with upland rice, following local practice.

Soil sampling

Four plots per site were laid out perpendicular to the slope avoiding gullies and isolated trees. Plots were large, 20 × 15 m, to avoid the variability linked to the meso relief, for example rills and termite moulds (Bigfoot & Inouye, 1988). To avoid autocorrelation between adjacent seedbank samples, a distance of 5 m was kept to account for large seed size and large mother plants (shrubs, trees). Litter was removed from the soil surface, and a cube of top soil 10 × 10 × 10 cm was collected every 5 m along diagonals, with the first sample a random distance from the edge, giving 10–12 samples per plot (Colbach et al., 2000). All soil samples from one plot were pooled. Sample depth was set at 10 cm following Oppong et al. (2003) and Witkowski and Wilson (2001), who recorded in the top soil layer of tropical no-till systems 93% (0–10 cm) and 99% (0–5 cm) of total seeds. Soil samples were air-dried and mixed thoroughly, and stones (<1% of volume), visible roots and other vegetative material (buds) were removed, and soil aggregates were gently broken. Two subsamples of 5 kg each were taken from every plot. Soil samples were kept dry in the dark. During storage, no visible signs of germination occurred. S10 and S6 were sampled before and immediately after burning. Post-burn soil cubes were taken next to the pre-burn samples. S5 and S0 were sampled immediately after burning.

Assessment of seeds in soil samples

The seedling-emergence method was used (Kropák, 1966) with 5 kg soil from one plot placed in each box (50 × 25 cm, 10 cm high). A total of 48 boxes were prepared, 24 in the sun and 24 in a shade nursery. Boxes receiving sunlight treatment were placed together and wrapped in netting to avoid seed contamination from outside. An identical nursery was built in a nearby bamboo groove with deep shade (shade treatment). Periodically, boxes were rerandomised. Samples were kept moist, and boxes were raised above the ground allowing drainage of excess water. Each set contained eight boxes of pre-burn soil samples from S10 and S6 (four plots per site) and 16 boxes post-burn samples from S10, S6, S5 and S0 (four plots per site). Emerging seedlings were counted, identified and removed. Unknown plants and less common species were counted and removed, while some individuals were left to grow up to the end of the experiment. These plants were identified as dried voucher specimens. Observations in the boxes were made every 3 or 4 days and continued over 140 days. In the last 20 days, no more seedlings appeared.

Field observations

Counting of seedlings

After burning, permanent quadrats of 0.5 × 0.5 m were installed centrally in every plot in all the sites except S10 (Table 1) to assess the fraction of the seedbank emerging in the field. Every 2 weeks throughout the cropping period, seedlings were counted, identified if possible, and removed by gently pulling them out. Four similar quadrats were added in S0 to evaluate the effect of the weeding tool on weed emergence. The sol surface was scraped each time the field was weeded, and the same observations as described above were made.


Relevés (Braun-Blanquet, 1964) were made before each weeding and at harvest from 3 × 3 m quadrats installed adjacent to the 0.5 × 0.5 m quadrats above. Data included height, cover and abundance per species. Each plant was determined to be either emerging from seed or resprouting from vegetative stocks. Observations were made on weed control by the farmer: number of passes, time and tool used. The purpose was to compare the floristic composition of the field with the seedbank.

Data analysis and statistics

The density of viable seeds (m−2) was calculated (after Albrecht, 2005) as:

  • display math

A single factor anova was conducted to determine the effect of site on total emergence, separately for seedbank (four sites) and field emergence (three sites) and to discriminate the effect of till and no-till on field emergence (one site), followed by mean comparison using the t-test. The Student's t-test of paired observations was used comparing emergence in sun and shade treatment. A two-factor anova was used to determine the significance of a response to burning (two sites).

The Spearman's rank correlation coefficient (rs) was used to investigate whether common species and functional groups emerged in the field proportionally to their abundance in the seedbank. Ranking species in order of rarity using absolute numbers gave many ties due to the high number of rare species; therefore, relative abundances were used. Because n > 10, the rs was tested as an ordinary product-moment correlation coefficient (t-value), following Sokal and Rohlf (1981).

The floristic similarity of seedbank and surface vegetation was analysed in a site ordination (Hill, 1979). All 28 floristic samples (16 seedbank records and 12 field relevés) were used. Abundance values in the sample-by-species matrix were transformed to a range of 0–4 (seedbank: 1 = 1–2 individuals, 2 = 3–5 individuals, 3 = 6–25 individuals, 4 ≥ 25 individuals; field: 1 = 1–2 individuals or <1% cover, 2 = 3–5 individuals or 1–5% cover, 3 = 5–25% cover, 4 ≥ 25% cover).


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Appendix 1: Species recorded in soil seedbank and corresponding surface vegetation with indication of abundance. Sites are four slash-and-burn rice fields in northern Laos. Field recordings were made in permanent quadrats from burning to harvest, 0 = no occurrence. The coppicing species are indicated: [r] = mainly resprouting sometimes seedling, [R] = only resprouting.

Emergence from the seedbank

The 6387 seedlings emerging from the soil samples belonged to 75 species, of which, 14 species, mainly woody and represented by few individuals, could not be identified down to species. The number of species per site was similar (24, NS). Annual and bi-annuals accounted for 2906 individuals, grasses for 305, woody plants for 292, Compositae seedlings for 3378 of which, 2819 seedlings were Chromolaena odorata, the most abundant species (see Appendix 1 for scientific names and life forms). Woody species, for example trees, Ficus spp., Kydia calycina, Paranephelium spirei, and climbers, for example Bauhinia ornata, Paederia pilifera, were more common in S10 (14 species) than in S6 (11 species) and in S5 and S0, both seven species. Cumulative emergence from S10, >4000 seedlings m−2, greatly outnumbered the other sites, around 2000 seedlings m−2 (= 0.007, Fig. 1A) due to a single species, the sprawling annual Mimosa diplotricha. This accounted for 68% of all seedlings in S10; 21% in S0, 12% in S5 and 0% in S6 (Fig. 1B). If Mimosa is omitted from the analysis, the emergence curves of the four sites are similar (data not shown) and total emergence across sites become similar (1750 seedlings m−2).


Figure 1. Cumulative seedling emergence from soil samples collected after burning in four sites (error bars represent standard error of the mean of total emergence, different letters against the error bars indicate statistical significance < 0.01). (A) Total seedlings, (B) Mimosa diplotricha seedlings, no Mimosa in S6.

Download figure to PowerPoint

The first seedlings emerged after 7–15 days of watering the soil samples and belonged to two species, M. diplotricha and Leucaena leucocephala. Subsequent emergence comprised mainly the annual species, then perennials and lastly long-lived woody species. The sequence was consistent across sites. Most species emerged concentrated into peak periods, each period dominated by a group of species (Table 2).

Table 2. Species by site table arrangement showing successive emergence from soil collected in four slash-and-burn sites shortly after burning (sunlit treatment)Thumbnail image of

Most species showed greater emergence in sunlight, including those species which continue their life cycle in shade, such as Panicum brevifolium (Table 3). Shade reduced emergence of most species, except Physalis angulata and Chromolaena odorata, although in fields, Chromolaena is a light-demanding weed and fallow species.

Table 3. Number of seedlings emerging in full sunlight vs. shade from soil samples of four sites (S0, S5, S6, S10)
  1. a

    Seedlings emerged from 2 × 120 kg of soil, densities converted to m2.

  2. b

    ns = non-significant, **< 0.01, ***< 0.001.

  3. c

    Degrees of freedom, d.f. 23 means that a species was present in each of the 24 paired boxes.

Total seedlingsa27162387ns23
Total species7027  
Annual and bi-annual herbs, total 30 species
Legum. Mimosoidae Mimosa diplotricha 78991**14
Tiliaceae Triumfetta rhomboidea 387132***20
Compositae Conyza sumatrensis 23234***23
Cyperaceae Cyperus laxus 9513**16
CompositaeBlumea ssp.580***18
Compositae Blumea balsamifera 442***20
Oxalidaceae Biophytum sensitivum 4011**14
Molluginaceae Mollugo pentaphylla 380***9
Euphorbiaceae Phyllanthus urinaria 3614***11
Compositae Ageratum conyzoides 292**8
Compositae Spilanthes paniculata 160***8
Legum. Caesalpinioidae Senna tora 118ns8
Rubiaceae Borreria laevis 60***4
Solanaceae Physalis angulata 463**12
Perennials, herbaceous and subwoody, total 25 species
Compositae Chromolaena odorata 4171866***23
Graminae Panicum brevifolium 1890***23
Commelinaceae Cyanotis cristata 260**9
Graminae Microstegium ciliatum 2313ns7
Graminae Saccharum spontaneum 190***14
Perennials, woody, total 20 species
Legum. Mimosoidae Leucaena leucocephala 11423**7
MoraceaeFicus sp. 2182**12
Ulmaceae Trema orientalis 70***6

The effect of burning on emergence was not consistent across sites; post-burn densities increased by 36% in S10 and reduced by 43% in S6 (Table 4). The number of species was not affected. Effects of burning on emergence were found to be positive in some species and negative in the majority of species, while a third group seemed unaffected (Table 4). Species stimulated by burning occurred in both sites and their abundance in the seedbank was reflected in high or low seedling densities after burning.

Table 4. Effect of burning on seedling emergence from soil samples collected before and after burning in two rice fields (S10, S6)
  1. a

    Seedlings emerged from 2 × 40 kg of soil, densities converted to m−2.

  2. b

    ns = non-significant, *< 0.05, **< 0.01.

  3. c

    Occurring in one site only.

Total seedlingsa m−2
Stimulated by burning
Mimosa diplotricha c 12982677ns 
Phyllanthus urinaria 4268ns**
Biophytum sensitivum 2789**
Cyanotis cristata c 2545ns 
Unaffected by burning
Conyza sumatrensis 218213nsns
Microstegium ciliatum c 6668ns 
Ageratum conyzoides 3845ns*
Ficus sp. 22719nsns
Saccharum spontaneum 1216nsns
Suppressed by burning
Chromolaena odorata 566279****
Leucaena leucocephala c 471208ns 
Triumfetta rhomboidea 409286ns**
Panicum brevifolium 199126nsns
Cyperus laxus 16463ns*
Blumea ssp.9861ns*
Blumea balsamifera 6633nsns
Senna tora c 4018ns 
Spilanthes paniculata c 4018ns 

Emergence in the field

A total of 124 species were identified in the rice fields S6, S5 and S0, of which 11 species occurred as seedlings and resprouting plants and 71 species occurred only as resprouting plants (Appendix 1). Typically, a 3 × 3 m quadrat contained 12 different coppicing species. The seedling densities in the fields ranged from 4 to 48 m−2 (Table 5). The sequence in which the various species emerged in the quadrats was similar to the order of germination from the soil samples. However, the species emerging after 51 days (Table 2) germinated in low numbers or were absent from the quadrats (Appendix 1). Thus, grasses and the subshrub Triumfetta rhomboidea did not emerge in the field during the observation period of 190 days, yet accounted for over 30% of emergence in the seedbank. These species are abundant in field margins and young fallows and may germinate after cropping.

Table 5. Correlation between seedling emergence from soil seedbank in samples taken shortly after burning cut vegetation and in the cultivated field plot from which samples were taken in four sites (S0, S5, S6, S10)
 Average density (m−2)CorrelationaComment
SeedbankbFieldc r s Effectdd.f.
  1. a

    Spearman rank correlation test.

  2. b

    Emergence in 5 kg soil converted to surface.

  3. c

    Cumulative emergences from burning to harvest in 1 m2.

  4. d

    ns = no correlation, *< 0.05 = significant agreement in abundance, (ns) = 0.05 < < 0.1, near-significant.

  5. e

    No emergence in one of the 12 field plots during rice cultivation.

Total seedlings18912600.495ns10 
Total species1460.327ns10 
Grasses2200No teste  >10% in seedbank
Chromalaena odorata 367100.589(ns)10 
Mimosa diplotricha 2571800.649(ns)6 
Triumfetta rhomboidea 4040No teste  >20% in seedbank

Comparison between seedbank and field

Site ordination (Fig. 2) of seedbank and surface vegetation shows separate clusters for seedbanks and field, probably because the field vegetation comprised mostly re-sprouting trees and lianas, which were either absent or occurred in low numbers in the seedbank. The compact cluster of the seedbank samples indicated a relatively small, uniform seedbank. It is noteworthy that the four plot scores from the same sites were not clustered, but were mixed with the plot scores of other sites, for seedbank records and field relevés alike. This demonstrates that the sites shared a common species pool.


Figure 2. Detrended correspondence analysis of soil seedbank (S0, S5, S6 and S10) and corresponding vegetation at sites (S0, S5 and S6), eigenvalues axis 1 0.587, axis 2 0.284. Open symbols represent vegetation relevés, closed symbols represent seedbank records, and each symbol corresponds to vegetation and seedbank for particular field.

Download figure to PowerPoint

The fraction of the seedbank emerging in the field varied according to site, 3%, 7% and 32% in S6, S5 and S0 respectively (Table 6). In a recent and frequently cultivated field (S0), a higher fraction of the seedbank emerged than in fields cultivated less frequently (S5 and S6). In S0, the majority of seedlings emerged early in the season, whereas in S5 and S6, most of the seedlings emerged late, outside the period of weed control. Superficial tillage in S0 triggered a double amount of seedlings to emerge: 76% of the seedbank or 1590 seedlings m−2 against 32% without soil disturbance (< 0.001). The effect was restricted to the first tillage/weeding and did not affect species composition.

Table 6. Seedlings emerging in three rice fields in relation to their soil seedbank: S0, S5 and S6 cultivated after 0, 5 and 6 years of fallow respectively
Total seedlings in field (m−2)63111446
Fraction of seedbank germinated32%7%3%
Period of emergence (m−2)
Between burning and sowing5082811
Weeding in rice993016
After last weeding245619
Major species emerging (m−2), with fraction of their specific–seedbank (%)
Mimosa diplotricha 516 (90%)24 (13%)1 (not in sb)
Chromolaena odorata 19 (10%)18 (3%)8 (2%)
Spilanthes paniculata 22 (83%)1 (7%)0 (not in sb)
Cyperus laxus 10 (17%)0 (not in sb)1 (3%)
Cyanotis cristata 5 (7%)0 (not in sb)7 (15%)

Total emergence from the seedbank samples was not correlated with densities in the corresponding field quadrats (Table 5). However, a positive correlation in abundance between seedbank and field flora was found in annual species (< 0.05; Table 5).


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Appendix 1: Species recorded in soil seedbank and corresponding surface vegetation with indication of abundance. Sites are four slash-and-burn rice fields in northern Laos. Field recordings were made in permanent quadrats from burning to harvest, 0 = no occurrence. The coppicing species are indicated: [r] = mainly resprouting sometimes seedling, [R] = only resprouting.

Characteristics of the seedbank

The largest variations in seedbank densities across sites were attributed to the invasive species M. diplotricha. The species, previously Mimosa invisa, was introduced in Indochina as a green manure (Poilane, 1952), but became a noxious weed in northern Thailand after its introduction for fencing in 1982 (Yimyam et al., 2007). Road-building machines transported seeds accidentally up to Luang Prabang in 1962. Our data suggest that the invasion of Mimosa over northern Laos is not complete. The geographic isolation of fields, high levels of forest cover, large seeds and high rates of predation among large seeds (Thompson, 1987) constitute barriers to spread and may explain the time lag between first arrival and infestation of fields.

The most abundant species in this study was Chromolaena odorata. Chromolaena was introduced in India in 1840, spread to Burma (1920), Thailand (1922), Laos (late 1920s) and invaded the shifting cultivation fields in northern Laos in the 1930s (Izikowitz, 2001). Farmers do not consider it a problematic weed due to slow juvenile growth (Roder et al., 1995a; Izikowitz, 2001; this study). Ageratum conyzoides was second after Chromolaena and first among the problem weeds in a frequency survey among 55 shifting cultivation fields in northern Laos, where it was negatively correlated with fallow age (Roder et al., 1995b). In this study, Ageratum scored 14th and 19th most common species in seedbank and field respectively, and no relation was found with fallow age.

Shade inhibited seed germination in most species. The preference of Chromolaena for germination in shady conditions, also recorded by Witkowski and Wilson (2001), may explain its dominance in young fallows. It is likely that in shifting cultivation systems, post-harvest canopy closure inhibits germination of most weeds and the addition of new seeds in the seedbank. Conversely, delayed canopy closure would allow prolonged reseeding, and this could affect the seedbank. The primary source of post-harvest overhead shade is regrowth from tree stumps, which were recorded in high density and diversity in this study.

Burning affects seeds in different ways. In the majority of species the heat destroyed up to about half of the seed stock, but a stimulant effect of fire on germination was found in a small group of weeds in accordance with Rico-Gray and Garcia-Franco (1992), who recorded twice as many seedlings emerging from post-burn soil samples (1815 m−2) than from pre-burn samples. If the seedbank is dominated by such species, total emergence after burning can exceed emergence without burning, otherwise seed emergence is strongly reduced after burning (de Rouw & van Oers, 1988; Hooper et al., 2005; de Mamade & de Araújo, 2008; this study). Heat stimulation of germination was reported under laboratory conditions in Mimosa (Chauhan & Johnson, 2008) and in other species (Benech-Arnold et al., 2000). Burning did not affect species composition of the seedbank, in contrast to a 40% or more reduction in species number found by Hooper et al. (2005) and de Mamade and de Araújo (2008). The conclusion that seeds from monocotyledons suffer more from fire than dicotyledons (Standish et al., 2007) is not confirmed in this study.

Dissimilarities of seedbank and field

We recorded twice as many species in the standing vegetation compared with the seedbank. The community of resprouting plants, mostly woody, was poorly represented in the seedbank, despite being well represented in the field vegetation, as also found in Yucatan (Rico-Gray & Garcia-Franco, 1992). The inverse was found by Chikoye and Ekeleme (2001), perhaps because species-poor Imperata fields were sampled which, usually lack woody populations.

Successive flushes of weeds emerged in fields, yet the first flush is responsible for the most severe competition with annual crops (Akobundo & Ekeleme, 2002; this study). Akobundo and Ekeleme (2002) suggested that later flushes replenish the seedbank as these weeds escape control and can set seed. Our data do not confirm this, as the seedbank was dominated by species of the first weed flush, the same group that causes greatest trouble to farmers, the target for most weeding and which maintained high levels of seed deposits in the soil. Many species emerging from the soil samples, after 50 days, comprising grasses and some Compositae, did not emerge in the field between burning and harvest, that is during 190 days of monitoring. These species account for 30–40% of the seedbank. Their abundance in the young fallow suggests that they will emerge after the harvest. More study is needed to identify the factor delaying their emergence in the cultivated field.

The small hand hoe, which is used for weeding on steep slopes in Indochina (Kunstadter & Chapman, 1978; Van Keer, 2003; Dupin et al., 2009), scrapes the soil surface rather than turning it. Hence, the tillage effect of bringing additional seeds to the surface where they are in a position to germinate, as well as burying seeds too deep for emergence, is largely absent. The almost threefold increase in weed emergence recorded after surface tillage supports farmers' decisions to only spot weeding as a means to limit competition.

Differences in weed populations among the rice fields appeared to be due to a larger fraction of the seedbank emerging during cropping. In a system without (deep) tillage, the recently dispersed seeds are mostly lying at the soil surface. During years of fallow, seeds have the opportunity of moving down to deeper soil layers via soil fauna activity and cracks. Superficially buried seeds are in a more favourable position to germinate than more deeply buried seeds. With frequent cultivation a greater fraction of the seedbank would be located in upper soil layers where it is in a better position to emerge. Shallow depth of burial should then be a crucial factor for high weed densities in fields subject to short-term fallowing. Profile studies are needed to prove this, together with the capacity of seeds to emerge from a certain depth of burial.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Appendix 1: Species recorded in soil seedbank and corresponding surface vegetation with indication of abundance. Sites are four slash-and-burn rice fields in northern Laos. Field recordings were made in permanent quadrats from burning to harvest, 0 = no occurrence. The coppicing species are indicated: [r] = mainly resprouting sometimes seedling, [R] = only resprouting.
  • Akobundo IO & Ekeleme F (2002) Weed seedbank characteristics of arable fields under different fallow management systems in the humid tropical zone of southeastern Nigeria. Agroforestry Systems 54, 161170.
  • Albrecht H (2005) Development of arable weed seedbanks during the 6 years after the change from conventional to organic farming. Weed Research 45, 339350.
  • Benech-Arnold R, Sánchez RA, Forcella F, Kruk BC & Ghersa CM (2000) Environmental control of dormancy in weed seed banks in soil. Field Crops Research 67, 105122.
  • Bigfoot DW & Inouye DW (1988) Spatial pattern analysis of seed banks: an improved method and optimized sampling. Ecology 69, 497507.
  • Braun-Blanquet J (1964) Pflanzensoziologie. Grundzüge der Vegetationskunde, 3rd edn. Springer-Verlag, Vienna.
  • Bech Bruun TB, Mertz O & Elberling B (2006) Linking yields of upland rice in shifting cultivation to fallow length and soil properties. Agriculture, Ecosystems and Environment 113, 139149.
  • Chauhan BS & Johnson DE (2008) Seed germination and seedling emergence of giant sensitiveplant (Mimosa invisa). Weed Science 56, 244248.
  • Chikoye D & Ekeleme F (2001) Weed flora and soil seedbanks in fields dominated by Imperata cylindrica in the moist savannah of West Africa. Weed Research 41, 475490.
  • Colbach N, Dessaint F & Forcella F (2000) Evaluating field-scale sampling methods for the estimation of mean plant densities of weeds. Weed Research 40, 411430.
  • Dupin B, de Rouw A, Phantahvong KB & Valentin C (2009) Assessment of tillage erosion rates on steep slopes in northern Laos. Soil & Tillage Research 103, 119126.
  • Garwood NC (1989) Tropical soil seed banks: a review. In: Ecology of Soil Seed Banks (eds MA Leck, VT Parker & RL Simpson), 149209. Academic Press, San Diego, USA.
  • Guevara S, Moreno-Casasola P & Sánchez-Ríos G (2005) Soil seed banks in the tropical agricultural fields of Los Tuxtlas, Mexico. Tropical Ecology 46, 219227.
  • Hill MO (1979) DECORANA - a FORTRAN program for Detrended Correspondence Analysis and Reciprocal Averaging. Cornell University, Ithaca, New York, USA.
  • Hooper E, Legendre P & Condit R (2005) Barriers to forest regeneration of deforested and abandoned land in Panama. Journal of Applied Ecology 42, 11651174.
  • Izikowitz KG (2001) Lamet hill peasants in French Indochina. Reprint White Lotus, Bangkok, Thailand.
  • Kellman MC (1974) The viable weed seed content of some tropical agricultural soils. Journal of Applied Ecology 11, 669677.
  • Kropák Z (1966) Estimation of weed seeds in arable soil. Pedobiologia 6, 105128.
  • Kunstadter P & Chapman EC (1978) Problems of shifting cultivation and economic development in Northern Thailand. In: Farmers in the Forest (eds P Kunstadter, EC Chapman & S Sabhasri), 222. East-West Centre, University Press Hawaii, Honolulu, USA.
  • Lestrelin G, Giordano M & Keohavong B (2005) When “Conservation” Leads to Land Degradation. Lessons from Ban Lak Sip, Laos. IWMI Research report 91. IWMI, Colombo, Sri Lanka.
  • de Mamade AM & de Araújo FS (2008) Effects of slash and burn practices on a soil seed bank of caatinga vegetation in Northeastern Brazil. Journal of Arid Environments 72, 458470.
  • MSEC (Management of Soil Erosion Consortium) (1999) Biophysical and Socio-Economic Inventories. Houay Pano Catchment KM-10 Village. Soil Survey and Land Classification Centre, Vientiane, Lao P.D.R.
  • Oppong FK, Osei-Bonsu K & Amoach FM (2003) The seed bank of weed species in three ecosystems in a semi-deciduous forest zone in Ghana. Tropical Science 43, 8085.
  • Plagge CE, Frolking S, Chini LP & Hurtt G (2008) Slash and Burn Agriculture: A Dynamic Spatio-Temporal Model of Shifting Cultivation Locations and Areas. American Geophysical Union, Fall meeting, abstract #B41B-0377.
  • Poilane E (1952) L'Eupatorium odoratum L. et d'autres plantes de couverture en Indochine. Revue International de Botanique Appliquée et Agriculture Tropicale 32, 496497.
  • Rico-Gray V & Garcia-Franco JG (1992) Vegetation and soil seed bank of successional stages in tropical lowland deciduous forest. Journal of Vegetation Science 3, 617624.
  • Roder W, Phengchanh S, Keoboulapha B & Maniphone S (1995a) Chromolaena odorata in slash-and-burn rice systems in Northern Laos. Agroforestry Systems 31, 7992.
  • Roder W, Phouaravanh B, Phengchanh S & Keoboulapha B (1995b) Relationships between soil, fallow period, weeds and rice yield in slash-and-burn systems of Laos. Plant and Soil 176, 2736.
  • Roder W, Phengchanh S & Keoboulapha B (1997) Weeds in slash-and-burn rice fields in Northern Laos. Weed Research 37, 111119.
  • de Rouw A & van Oers C (1988) Seeds in a rain forest soil and their relation with shifting cultivation. Weed Research 28, 373381.
  • de Rouw A, Soulileuth B, Phanthavong K & Dupin B (2005) The adaptation of upland rice cropping to ever-shorter fallow periods and its limit. In: Poverty reduction and shifting cultivation stabilization in the uplands of Lao PDR (eds B Bouahom, A Glendinning, S Nilsson & M Victor), 139148. NAFRI, NAFES, NUOL, Vientiane, Lao P.D.R.
  • Saito K, Linquist B, Keobualapha B, Phanthaboon K, Shiraiwa T & Horie T (2006) Cropping intensity and rainfall effects on upland rice yields in northern Laos. Plant and Soil 284, 175185.
  • Sokal RR & Rohlf FJ (1981) Biometry. The Principles and Practice of Statistics in Biological Research, 2nd edn. W.H. Freeman and Company, New York, USA.
  • Standish RJ, Cramer VA, Wild SL & Hobbs RJ (2007) Seed dispersal and recruitment limitation are barriers to native recolonization of old-fields in western Australia. Journal of Applied Ecology 44, 435445.
  • Thompson K (1987) Seeds and seed banks. New Phytologist 106 (Suppl.), 2334.
  • Van Keer K (2003) On-Farm Agronomic Diagnosis of Transitional Upland Rice Swidden Cropping Systems in Northern Thailand. PhD thesis, Katholieke Universiteit Leuven, Belgium.
  • Witkowski ETF & Wilson M (2001) Changes in density, biomass, seed production and soil seed banks of the non-native invasive plant, Chromolaena odorata, along a 15 year chronosequence. Plant Ecology 152, 113.
  • Yimyam N, Rerkasem B & Rerkasem K (2007) Shifting cultivation in northern Thailand: land use changes and alternative management for sustainable rural livelihoods. In: Shifting Cultivation in Asia (eds KG Saxena, L Luohui & K Rerkasem), 309319. M/s Bishen Singh Mahendra Pal Singh, Dehra Dun, India.

Appendix 1: Species recorded in soil seedbank and corresponding surface vegetation with indication of abundance. Sites are four slash-and-burn rice fields in northern Laos. Field recordings were made in permanent quadrats from burning to harvest, 0 = no occurrence. The coppicing species are indicated: [r] = mainly resprouting sometimes seedling, [R] = only resprouting.

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Appendix 1: Species recorded in soil seedbank and corresponding surface vegetation with indication of abundance. Sites are four slash-and-burn rice fields in northern Laos. Field recordings were made in permanent quadrats from burning to harvest, 0 = no occurrence. The coppicing species are indicated: [r] = mainly resprouting sometimes seedling, [R] = only resprouting.
 SeedbankaRice fieldsb
  1. a

    Code for abundance (m−2) in seedbank: + = 1–10 seedlings, ++ = 11–100 seedlings, +++ > 101-seedlings.

  2. b

    Code abundance in field: + = <1% cover, ++ = 1–10% cover, +++ = >10% cover.

  3. c

    Small seedlings of Blumea mollis, B. lacera and B. membranacea could not be distinguished.

  4. d

    Only germinating in the shade.

Broad-leaved herbs, annuals and bi-annuals
CompositaeAgeratum conyzoides L+++++
CompositaeBidens pilosa L.0+
CompositaeBlumea balsamifera (L.) DC.+++
CompositaeBlumea lacera (Roxb.) DC.(++)c+
CompositaeBlumea membranacea DC.(++)c+
CompositaeBlumea mollis (D.Don) Merr.(++)c+
OxalidaceaeBiophytum sensitivum (L.) DC.+++
RubiaceaeBorreria laevis (Lmk.) Griseb.+0
AmarantaceaeCelosia argentea L.+0
CompositaeConyza sumatrensis (Bth.) Walk.+++++
TiliaceaeCorchorus aestuans L.+++
CompositaeCrassocephalum crepidioides S.Moore+++
UrticaceaeDistemon indicum Wedd.+0
EuphorbiaceaeEuphorbia heterophylla L.0+
EuphorbiaceaeEuphorbia hirta L.+++
RubiaceaeHedyotis ovatifolia Cav.+0
SaururaceaeHouttuynia cordata Thunb.+0
ConvolvulaceaeJacquemontia paniculata (Burm.f.) Hall.f. var. paniculata+0
CompositaeLactuca indica L.+0
OnagraceaeLudwigia hyssopifolia (G.Don) Exell++
LeguminosaeMimosa diplotricha C.Wright ex Sauv. var. diplotricha++++++
MolluginaceaeMollugo pentaphylla L.++++
CucurbitaceaeMomordica charantia L.0+
CucurbitaceaeMukia javanica (Miq.) C.Jeffrey0+
OxalidaceaeOxalis corniculata L.0+
PassifloraceaePassiflora foetida L.0+
EuphorbiaceaePhyllanthus amarus Schumach. & Thonn.0+
EuphorbiaceaePhyllanthus urinaria L.+++
EuphorbiaceaePhyllanthus spp+0
SolanaceaePhysalis angulata L.++
LeguminosaeSenna tora Roxb.+0
CompositaeSpilanthes paniculata Wall.++++
SchrophulariaceaeTorenia cordifolia Benth.0+
TiliaceaeTriumfetta rhomboidea L.++++
Grasses and sedges, mostly perennials
GraminaeCentotheca lappacea (L.) Desv. var. lappacead+0
CyperaceaeCyperus laxus Lam.+++

Imperata cylindrica (L.) P.Beauv. var. major (Nees)

C.E. Hubb. ex Hubb. & Vaugh.

0+ [R]
GraminaeMicrostegium ciliatum A.Camus+++ [R]
GraminaePanicum brevifolium L.+++0
GraminaeSaccharum spontaneum L.++++ [r]
GraminaeThysanolaena maxima Kuntze0+++ [R]
Broad-leaved perennial herbs
AraceaeAmorphophallus paeoniifolius (Dennst.) Nicolson0+
LoganiaceaeBuddleia asiatica Lour.+0
CommelinaceaeCommelina benghalensis L.++++ [r]
CommelinaceaeCyanotis cristata (L.) D. Don+++++ [r]
LeguminosaeDesmodium gangeticum (L.) DC.++
UrticaceaeElatostema monandrum (Ham. ex D.Don) Harad+0
AmarantaceaeGomphostemma strobilinum Wall. ex Bth.+0
RubiaceaeHedyotis coronaria (Kurz) Sweet var. pubescens Kurz+0
LeeaceaeLeea indica (Burm.f.) Merr.0+ [R]
LabiataePlectranthus sp. 1++0
SolanaceaeSolanum torvum Sw.+0
BoraginaceaeTournefortia sp.0+ [R]
Perennial vines and subwoody climbers
PassifloraceaeAdenia heterophylla (Bl.) Koord. ssp. heterophyllad+0
LeguminosaeCajanus goensis Dalzell0+ [R]
CompositaeChromolaena odorata (L.) R. King & H. Robinson++++++ [r]
MenispermaceaeCyclea barbata Miers0+ [R]
DioscoreaceaeDioscorea alata L.++++ [R]
DioscoreaceaeDioscorea arachidna Prain & Burkill var. arachidna0+ [R]
DioscoreaceaeDioscorea pentaphylla L.0+ [R]
ConvolvulaceaeIpomoea triloba L.+0
ConvolvulaceaeLepistemon binectariferum Kuntze0+ [R]
RubiaceaeMussaenda sp. 1+0
RubiaceaeOphiorrhiza hispidula Wall. ex G. Don var. hispidula+0
RubiaceaePaederia pilifera Hk.f.++++ [R]
PassifloraceaePassiflora siamica Craib0+
SmilacaceaeSmilax ovalifolia Roxb.+0
MenispermaceaeStephania crebra For.0+ [R]
VitaceaeTetrastigma laoticum Gagnep0+ [r]
VitaceaeTetrastigma sp.0+ [R]
AcanthaceaeThunbergia grandiflora Roxb.++++ [R]
AsclepiadaceaeZygostelma benthami Baill.0++ [R]
UrticaceaeBoehmeria zollingeriana Wedd. var. zollingeriana0+ [R]
AsclepiadaceaeCapparis sabiaefolia Hook.f. & Thomson0+ [R]
AsclepiadaceaeCapparis zeylanica Wight & Arn.0++ [r]
LeguminosaeCrotalaria dubia Grah. ex Bth.0+
FlacourtiaceaeFlacourtia indica (Burm.f.) Merr.0+ [R]
SterculiaceaeHelicteres elongata Wall.0+ [R]
SapindaceaeLepisanthes rubiginosa (Roxb.) Leenh.0++ [R]
MyrsinaceaeMaesa ramentacea Wall.0++ [R]
EuphorbiaceaeSauropus quadrangularis (Willd.) M.-A.+0
LeguminosaeSenna hirsuta (L.) Irwin & Barneby var. hirsuta+0
Large woody climbers
LeguminosaeAcacia concinna Wall.0++ [R]
ApocynaceaeAganosema marginata (Roxb.) G. Don0+ [R]
ApocynaceaeAmalocalyx microlobus Pierre0+++ [R]
LeguminosaeBauhinia ornata Kurz var. kerrii (Gagnep.) K. & S.S. Lar.d+0
LeguminosaeCaesalpinia decapetala (Roth) Alst.++ [R]
CelastraceaeCelastrus paniculatus Willd.0+ [R]
CombretaceaeCombretum decandrum Roxb.0+ [R]
CombretaceaeCombretum pilosum Roxb.0++ [R]
CombretaceaeCombretum sp.0++ [R]
LeguminosaeDiphyllarium mekongense Gagnep.0++ [R]
CelastraceaeEuonymus cochinchinensis Pierred++ [R]
OleaceaeJasminum nervosum Lour.0+ [R]
LeguminosaeMillettia pachycarpa Benth.0+ [R]
CelastraceaeReissantia indica (Willd.) Halle0+ [R]
ConnaraceaeRourea minor (Gaertn.) Leenh. ssp. Minor0+ [R]
MenispermaceaeTiliacora triandra Diels0+ [R]
LeguminosaeAdenanthera microsperma Teijsm. & Binn.0+ [R]
LeguminosaeAlbizia odoratissima (L.f.) Benth.0++ [R]
EuphorbiaceaeAntidesma acidum Retz.0++ [R]
EuphorbiaceaeAntidesma sootepense Craib0++ [R]
CombretaceaeAnogeissus acuminata (Roxb. ex DC.) Guill. & Perr.0++ [R]
EuphorbiaceaeApurosa octandra (Buch.-Ham. ex D.Don) A.R.Vickery0+++ [R]
MoraceaeArtocarpus lakoocha Roxb.0+ [R]
Euphorbiaceae Baccaurea cf siamensis 0+ [R]
EuphorbiaceaeBridelia stipularis Blume0++ [R]
EuphorbiaceaeBridelia tomentosa Bl.0+ [R]
LeguminosaeCallerya atropurpurea (Wall.) Schot var. atropurpurea0+ [R]

Cratoxylum formosum (Jack) Dyer ssp parviflorum

(Kurz) Gog.

0+++ [R]
AnnonaceaeEllipeia dulcis (Dun.) C. Meade0+ [R]
MyrtaceaeEugenia cumini (L.) Druce0+ [R]
MyrtaceaeEugenia syzygioides (Miq.) Hend.0+ [R]
MyrtaceaeEugenia sp.0+ [R]
BignoniaceaeFernandoa adenophylla (Wall. ex G.Don) Steenis0++ [R]
MoraceaeFicus hispida L.f.var. hispida0+ [R]
MoraceaeFicus sp. 1+0
MoraceaeFicus sp. 2++++ [R]
SterculiaceaeFirmiana colorata (Roxb.) R.Brown0++ [R]
SimaroubaceaeHarrisonia perforata Merr.0++ [R]
ApocynaceaeHolarrhena pubescens Wall. & G.Don0+ [R]
EuphorbiaceaeKydia calycina Roxb.++ [R]
LythiraceaeLagerstroemia macrocarpa Wall. var. macrocarpa0+ [R]
LythiraceaeLagerstroemia tomentosa Presl0+ [R]
LeguminosaeLeucaena leucocephala (Lmk.) De Wit++++ [R]
EuphorbiaceaeMallotus philippensis (Lam.) Mull.Arg.0++ [r]
MoraceaeMorus alba L.++ [R]
BignoniaceaeOroxylum indicum (L.) Kurz0++ [r]
SapindaceaeParanephelium spirei Lecomte+++ [R]
LauraceaePhoebe lanceolata Nees0++ [R]
BurseraceaeProtium serratum Engl.0+ [r]
SterculiaceaeSterculia lanceolata Cav. var. lanceolata0+ [R]
UlmaceaeTrema orientalis (L) Bl.+++
VerbenaceaeVitex canescens Kurz0+ [r]
CycaceaeCycas tonkinensis L. Lind. & Rod.0+ [r]
SchizaeaceaeLygodium flexuosum (L.) Sw.0+++ [R]
GraminaeBambusa tulda Roxb.0+++ [R]
PalmaeCalamus viminalis Willd.0++ [R]
OphioglossaceaeOphioglossum petiolatum Hk.0+ [R]