1. An early inhibition of germination in seeds of Silene gallica and Brassica campestris which were continuously exposed to the light environment under an establishing wheat canopy, was observed in two different experiments. Inhibition occurred c.15 days after crop emergence, when the canopy leaf area index (LAI) was below one and the red (R):far-red (FR) ratio recorded under the canopy was well above 0·8.
2. This inhibitory effect was either overcome by filtering FR light through a solution of CuSO4 or could be artificially imposed by simulating the canopy with filters yeilding a R:FR ratio of 0·95 and 0·8. These results show that light subtly enriched with FR was the environmental factor regulating germination below the developing canopy.
3. Exposure to canopy-filtered light pulses of 1 h (presumably sufficient to saturate a low fluence response, LFR) did not inhibit seed germination. Moreover, such treatment promotes germination up to an extent similar to that previously observed in the laboratory after a saturating pulse of R light. Instead, prolonged exposures were required to inhibit germination. These results, together with the relatively high R:FR ratios measured below the canopy in early stages of its establishment, suggest that a high irradiance response (HIR) would be involved in such a regulation.
4. This capacity to detect small environmental light-quality modifications when exposed to high irradiances, would allow the seeds from these species to detect the presence of a canopy in the very early stages of its establishment and to stay in ‘safe’ pre-germination phases when the probability of successful seedling establishment is low.
The presence of a leaf canopy reduces soil thermal amplitude, thus preventing germination of species with seeds which require temperature fluctuations to be released from dormancy (Thompson & Grime 1983). Nevertheless, pronounced modifications in soil thermal amplitude, large enough to reduce seed germination, can be expected with light interception values of 50% or higher, a condition that can only be imposed by a relatively dense canopy (Benech-Arnold et al. 1988).
The presence of a leaf canopy also reduces the photon flux density (PFD) and produces spectral changes in the light environment, mainly through the reduction of the R:FR ratio, resulting from the strong absorption by chlorophyll in the photosynthetic part of the spectrum (400–700 nm) (Smith 1982; Pons 1992). These spectral variations in the light environment can be perceived by the seeds on the soil surface through photoreceptors, particularly those from the phytochrome family, establishing an amount of Pfr (the active phytochrome form for germination). Depending on whether the established value is above or below a Pfr:Ptotal threshold value (which depends on the species and on the degree of dormancy of the seeds), germination would be either promoted or inhibited, respectively (Taylorson & Borthwick 1969; Ellis, Hong & Roberts 1989; Pons 1992; Bewley & Black 1994). Within this context, Frankland & Poo (1980) showed that the R:FR ratio which had to be imposed by a Sinapis alba canopy to establish a Pfr:Ptotal value capable of inhibiting germination of Plantago major seeds, must be at least 0·6. To reach those R:FR ratio values, the canopy LAI must be higher than 1·4. In agreement with those results, Deregibus et al. (1994) found a threshold of R:FR ratios of 0·5–0·8 below which germination of Lolium multiflorum seeds located at the canopy base would be inhibited. Again, these inhibitory R:FR ratios can only be imposed by a relatively dense canopy. Because marked changes in the thermal and light environment do not occur before the canopy reaches high LAI values, it has been assumed that seeds cannot detect the presence of an establishing canopy (i.e. a crop canopy) (Fenner 1980; Benech-Arnold et al. 1988). In other words, to our knowledge, no work has focused on evaluating whether an anticipation of a few days in the emergence of one species can inhibit the germination of seeds from other species.
In this study we show that the seeds of some weeds (Brassica campestris L. and Silene gallica L.) can detect the presence of a canopy during the very early stages of its establishment (LAI < 1), by perception of slight reductions in the R:FR ratio of natural light. This reduction occurs at an early stage of canopy development at high irradiances.
We also investigated which mode of action of phytochrome could be involved in this phenomenon.
Materials and methods
Seeds of six common farmland weeds in the Argentinean Pampas were used: Sonchus oleraceus L., Brassica campestris L., Silene gallica L., Chenopodium album L., Galinsoga parviflora Cav. and Amaranthus hybrydus L. var. quitensis (H.B.K.) Cov.
The seeds were hand-collected from plants growing in a wheat field at Balcarce (37°45′S, 58°15 ′W), Argentina, during 1996 and 1997, and stored in glass jars at room temperature.
In order to characterize the light response of seeds from the different species, some preliminary tests were performed in the laboratory before field experiments started. Fifty seeds from each species (except Ch. album and S. oleraceus) were sown in Petri dishes (9 cm diameter) on a thin cotton-wool layer and two layers of filter paper (Whatman No. 3), moistened with distilled water and placed in darkness for 24 h to allow imbibition. The seeds were exposed for seven consecutive days to different light treatments: (1) 2 min of R light per day; (2) 2 min of R light, followed by 15 min of FR light per day; (3) 20 min of R light per day; (4) 20 min of R light, followed by 15 min of FR light per day; (5) continuous darkness. Light sources used to give R and FR pulses are described in Casal et al. (1991). During the experiment, the Petri dishes were maintained in dark incubators with temperatures previously reported to be optimal for the germination of each species (S. gallica 15 °C, G. parviflora 25 °C, A. quitensis 30 °C and B. campestris 20/25 °C). Visible radicle protrusion was the criterion for germination recordings.
Three experiments were performed in the experimental field of the Facultad de Agronomía (Universidad de Buenos Aires, Argentina) (34°25′S, 58°25′W), on three consecutive years: August 1996, experiment no. 1; August 1997, experiment no. 2; October 1998, experiment no. 3.
Experiment no. 1
The objective of this exploratory experiment was to determine whether the modifications in the light environment produced by a wheat canopy during early stages of its establishment were enough to regulate the germination of various weed seeds which were located at the soil surface. Therefore, wheat plots (0·8 m × 0·45 m, three interrows) were sown, establishing a density of 350 plants m−2, with a distance of 15 cm between rows (north–south row orientation). After crop sowing, the first 5 cm of the interrows soil was replaced with soil that had been oven-sterilized at 90 °C for 7 days. At crop emergence (visible first leaf tip) 100 seeds of each species, were randomly distributed on the soil surface of each interrow (300 seeds per plot).
Four different treatments were performed: (1) environment imposed by the presence of the establishing crop canopy; (2) modification of the light environment R:FR ratio, by removing FR with filters filled with a solution of CuSO4; (3) bare soil; (4) as treatment (1) but with filters filled with distilled water (control treatment).
R:FR ratio modification under the establishing canopy, was carried out by filtering light through a CuSO4 solution contained in polycarbonate alveolar plaques. CuSO4 is known to retain a great proportion of the FR radiation, without affecting photosynthetic active radiation (PAR) (Casal & Sánchez 1994). The CuSO4 solution was calibrated to obtain a 1·2 R:FR ratio, using concentration and procedures as described by Ballaré, Scopel & Sánchez (1991).
Experiment no. 2
The objective of this experiment was to describe in detail the light and thermal environment that prevailed at the soil surface during the first 30 days after crop emergence. We also varied the length of the period of exposure of seeds to canopy-filtered light, in order to establish whether the seed responses to the presence of an establishing canopy observed in experiment no. 1, can be saturated with a relatively short exposure or if, in contrast, long exposures were required to trigger such a response. Responses to different durations of exposure would reveal, at least in a preliminary way, the phytochrome mechanism involved in this regulation. For this experiment, wheat plots (3 m × 1·15 m) were sown, establishing a density of 250 plants m−2, with a distance of 15 cm between rows (north–south row orientation).
Groups of 35 seeds of each of the species used in experiment no. 1 (except S. oleraceus and G. parviflora), were sown on the surface of Petri dishes containing sterilized soil as substrate (see experiment no. 1 for sterilizing procedure). The lids of the dishes were covered with aluminium foil to prevent light reaching the seeds. The substrate was moistened with distilled water and placed in the dark for 24 h to allow seed imbibition.
At crop emergence (visible first leaf tip) the dishes, covered with their lids, were randomly distributed and buried between rows within the canopy plots and in bare-soil plots. They were buried in such a way as to leave the seeds at the soil-surface level.
Four different treatments were established: (1) continuous exposure to the light environment generated by the presence of the crop canopy, the dishes remained uncovered permanently; (2) exposure to 1 h canopy-filtered light pulses, the dishes were uncovered for 1 h at midday, remaining covered the rest of the day; (3) continuous dark, the dishes remained constantly covered; (4) continuous exposure to bare-soil light environment, the dishes remained permanently uncovered in a plot without canopy.
Experiment no. 3
In this experiment we artificially reproduced the light environment imposed by the presence of an establishing canopy by filtering sunlight through acetates filters.
Groups of 50 seeds of B. campestris and S. gallica were conditioned as in experiment no. 2, although the lids of the dishes were not covered with aluminium foil.
Filters were made by combining violet acetate sheet (La Casa del Acrílico, Buenos Aires, Argentina), in the form of strips of different width, over a transparent acetate mounted on a square metal wire structure. The filters were calibrated in order to establish three different R:FR ratios: (1) neutral filter (N) (made entirely with transparent acetate, which did not affect natural daylight R:FR ratio); (2) FR, a R:FR of 0·95; (3) FR +, a R:FR of 0·8. Transmission of the acetate material in a 250–750 wavelength band, was measured with an spectrophotometer (Genesis II, Spectronic Inc., Rochester, NY, USA) in order to ensure that the filters did not seriously affect PAR and other spectral bands except those corresponding to R and FR wavelengths (Fig. 5 inset). A translucent plastic diffuser was added to each filter to homogenize the incident light radiation and the R:FR ratio under the filters.
Four filters of each type (N, FR and FR +) were placed in bare ground, orientated towards the sun's rays. One dish with seeds of each species was buried below each filter, in a similar way as described in experiment no. 2. The dishes were taken to the field covered with black plastic bags, in order to prevent light from reaching the seeds before they were buried below the filters.
In the three experiments the following determinations were carried out in all treatments: (1) temperature measured at soil surface level using sensors (Li-Cor 1015, Li-Cor, Inc., Lincoln, NE, USA) connected to a data logger (Li-Cor 1000, Li-Cor, Inc., Lincoln, NE, USA); (2) R:FR ratio at soil level, was measured with a Skye sensor (SKR 100, Llandrindod Wells, UK). R:FR ratio was measured at midday in experiment no. 1 and at three different times of the day in experiment nos 2 and 3: in the morning (09.00 h), midday (12.00 h) and in the afternoon (17.00 h); (3) weed emergence was monitored. In experiment nos 1 and 2, non-destructive measurements of the crop LAI were also recorded with a Li-Cor LI2000 Canopy Analyser (Li-Cor, Inc., Lincoln, NE, USA).
The plots were watered at least five times a day, in order to maintain a constant moisture content in the soil surrounding the seeds.
Data were analysed using a general anova analysis (P < 0·05) and a Tukey's Studentized Range (HSD) Test (P < 0·05) for mean comparison. Germination data were arcsine-transformed for the statistical analysis (Sokal & Rohlf 1969) but are presented as percentages.
Preliminary tests revealed phytochrome action in the low fluence response (LFR) in seed germination of all the species through the photoreversibility induced by short pulses of R and FR light (Casal & Sánchez 1998) (Table 1). However, in B. campestris seeds, the promotive effect exerted by exposure to 20 min of R light could not be reverted by subsequent exposure to 15 min of FR, suggesting that scape time for phytochrome action in this species is shorter than 20 min. All the species showed low or null germination percentages under continuous darkness.
Table 1. Germination percentages of seeds exposed to different light treatments in the red (R) and far-red (FR) bands. Seeds were imbibed in darkness for 24 h in Petri dishes and were exposed for 1 week to 2 min of R light per day (2 R), 2 min of R light followed by 15 min of FR light per day (2 R + 15 FR), 20 min of R light per day (20 R), 20 min of R light followed by 15 min of FR light per day (20 R + 15 FR) or maintained under continuous darkness (Dark). After treatments were given the dishes were maintained in dark incubators with temperatures previously reported to be optimal for the germination of each species and germination was recorded until no more germinated seeds were observed. Criterion of germination was visible radicle protrusion. Different letters indicate significative differences between treatments (Tukey's test, P < 0·05)
2 R + 15 FR
20 R + 15 FR
Experiment no. 1
Thirty days after crop emergence germination of all weed species in all treatments had almost finished, except for A. quitensis and G. parviflora, which showed later germination flushes (Fig. 1). A significantly higher number of B. campestris and S. gallica seedlings emerged at bare soil as compared to the number emerged under the establishing canopy (Fig. 1). Those differences were already established as early as 15 days after the crop emergence. Although no LAI and R:FR were recorded at this time, 5 days later (20 days after crop emergence) the R:FR ratio recorded under the canopy (0·9) was slightly lower than that recorded at bare soil (1·15) and the crop canopy presented a LAI of 1·2. The rest of the species studied did not show significant differences in germination percentage between seeds placed on bare soil or under the canopy in early stages of crop establishment. In those species for which clear early differences between bare soil and canopy presence were observed (B. campestris and S. gallica), filtering FR in the interrows using CuSO4 solution produced a similar amount of emergence to that recorded on bare soil, thus overcoming most of the inhibition exerted by the presence of the canopy (Fig. 1). In contrast, polycarbonate plaques containing distilled water did not produce the same effect, suggesting that this regulation is likely to be exerted through changes in the R:FR ratio of the light that reaches the soil surface. Although, germination in B. campestris below the filters was lower than that observed under the canopy and bare soil, probably owing to an influence of the filters on other environmental variables, inhibition of germination was clearly overcome. In the other species, the artificial modification of the light environment did not significantly affect seed germination.
It should be noted that germination of B. campestris and S. gallica on bare soil was similar to that obtained in the laboratory after a saturating R light pulse. In contrast, in the other species, where no canopy regulation was observed, lower germination than that obtained in the laboratory after a R light pulse was recorded at bare soil, suggesting that germination of these species could be inhibited by continuous exposure to direct sunlight.
Experiment no. 2
The evolution of the LAI (Fig. 2), thermal and light environment (Fig. 2), were monitored to characterize the environment prevailing under the canopy at very early stages of crop establishment.
The range of daily average temperatures recorded during the experimental period was 15–28 °C, and extreme maximum and minimum values of 30·4 °C and 13·5 °C, respectively, were registered. No differences in the thermal environment were found between the different treatments (data not shown).
To analyse the germination dynamics of the different species, they were grouped by similar responses to the treatments.
Species that detected the presence of an establishing canopy
A similar pattern of response was observed in B. campestris and S. gallica seeds (Fig. 3). Forty days after crop emergence the germination percentage observed in seeds exposed to continuous canopy-filtered light was much lower than that registered for seeds placed on bare soil. However, seeds that received pulses of only 1 h of canopy-filtered light were not inhibited and produced similar germination percentages to those observed at bare soil. On the other hand, the germination percentages recorded with seeds that had been held constantly in the dark were similar to those observed on bare soil for B. campestris, while almost no germination was observed with S. gallica seeds.
For B. campestris differences in germination percentages between seeds that had received pulses and those that had been exposed to continuous canopy-filtered light, were established as early as 10 days after crop emergence (Fig. 3a). At that moment the crop LAI was below 0·2 (Fig. 2) and the R:FR ratio registered in the interrows at the three different times of day was never below one, although it was significantly different from the value observed at bare soil (Fig. 2). For, S. gallica differences were established 17 days after crop emergence (Fig. 3b), when the crop LAI was below 0·3 (Fig. 2) and the R:FR ratio recorded was always above 0·9 (Fig. 2). The results from this and the previous experiment show that, for these two species: (1) direct solar radiation given in long exposures does not inhibit germination, unless it is, at least, subtly enriched with far-red and (2) only continuous exposure to canopy-filtered light inhibited seed germination. It was noticeable that B. campestris seeds in continuous darkness in the field, showed high germination percentages (Fig. 3a), while those tested in the dark in the laboratory presented low germination percentages (Table 1). Field temperatures were higher than those used in the laboratory experiments and this possibly overcame the light requirement of these seeds.
Species that did not detect the presence of an establishing canopy
Forty days after crop emergence Ch. album and A. quitensis seeds that were either exposed to continuous canopy-filtered light or those placed on bare soil, presented germination percentages that were lower than 10% (Fig. 4). However, seeds exposed to 1 h canopy-filtered light pulses presented germination percentages that were much higher than those observed in the others three treatments, although none of the species showed germination percentages higher than 40%.
These results, together with those observed in experiment no. 1, suggest that in these species germination is inhibited by prolonged exposure to direct sunlight, in contrast to the first group of species.
Eexperiment no. 3
As in experiment no. 2, no differences in thermal environment were found between the different treatments (data not shown). The range of daily average temperatures recorded during the experimental period was 13–30 °C, with extreme maximum and minimum values of 33·8 °C and 10 °C, respectively.
Subtle changes in the R:FR ratio as a result of filtering sunlight in the 500–650 nm band [FR (R:FR 0·95) and FR + (R:FR 0·8)] inhibited germination of seeds of both species in relation to what occurred below the neutral filter (R:FR 1·07) (Fig. 5). This inhibition was particularly intense in S. gallica seeds and matched that observed in previous experiments. However, no significant differences between germination percentages under filters FR and FR + was observed for the two species. In contrast, inhibition of germination of B. campestris seeds placed below FR and FR + filters was not as severe as observed in experiment nos 1 and 2. This could be the result of the higher temperatures that prevailed in this experiment, which would probably allow many seeds of this species to escape from the light-imposed inhibition.
Several authors have demonstrated that seeds located at the soil surface can detect dense canopy and avoid futile germination by sensing changes produced in the light quality (Taylorson & Borthwick 1969; Górski 1975; Fenner 1980; Silvertown 1980; Bewley & Black 1994). Until now, it was assumed that this perception could occur only when a relatively high LAI value is reached (lowering R:FR sufficiently to establish a Pfr:Ptotal that could inhibit germination). However, the present results show that seeds of some species (B. campestris and S. gallica) are also able to detect the presence of an incipient canopy (with LAI values < 1) through the perception of subtle changes in the light environment.
The first evidence supporting this proposition was found in experiment no. 1, where an early inhibition of seed germination under an establishing canopy was observed, in relation to the higher number of seeds germinated at bare soil (Fig. 1). This regulation took place at very early stages of crop establishment, when the canopy LAI was c.1 and the R:FR ratio was close to 0·8. No modifications in the thermal environment were detected at this stage; however, subtle undetected changes in soil temperature cannot be ruled out as responsible for changing the germination response of the seeds. Because filtering far-red from the light environment using CuSO4 filters overcame the inhibitory effect imposed by the establishing canopy, the proposition that the inhibitory effect was exerted through subtle modifications in the light environment is a plausible one. This proposition was further supported by the results from experiment nos 2 and 3. Experiment no. 2, showed that differences in terms of germination between seeds exposed for 1 h and continuously to the light environment prevailing under the wheat canopy, were already established c.15 days after crop emergence (Fig. 3). At that time, the canopy LAI was around 0·3 (Fig. 2) and the R:FR ratio recorded under the canopy at any time during the day was above 0·9 (Fig. 2). This demonstrates that, although subtle FR enrichment is needed to inhibit germination, prolonged exposures are also required; indeed, 1 h pulses did not produce inhibition but, rather, promoted germination as compared with dark controls (Fig. 3). Finally, experiment no. 3 showed that using filters, a R:FR ratio similar to that prevailing below the wheat canopy in those early stages (0·95 and 0·8), also produced inhibition of germination of these species when seeds were continuously exposed to the light environment (Fig. 5). In the case of S. gallica, this inhibition was like that observed in experiment nos 1 and 2 under real canopies. Taken together, the results of the three experiments demonstrate that the inhibition of germination of seeds subjected to continuous exposure to canopy-filtered light, takes place during very early stages of crop establishment and that to be inhibitory, sunlight has to be at least subtly enriched in FR.
The R:FR ratios measured below the canopy during the early phase of its establishment, would establish Pfr:Ptotal values which would be presumably well above the threshold for promoting seed germination through the low irradiance response system (LFR), as reported for other species (Frankland & Poo 1980; Deregibus et al. 1994). Indeed, the germination percentages observed when canopy-filtered light was given in pulses of 1 h (presumably sufficiently to saturate an LFR) were as high as those observed as a result of the promotive effect of a saturating pulse of red light in the laboratory. Consequently, such R:FR ratios do not provide an explanation for the inhibition observed in the seeds exposed continuously to the canopy-filtered light, based only in a low irradiance system mediated response.
Górski & Górska (1979), working with lettuce seeds, determined that as irradiance increases, the inhibitory effects exerted by the far-red light components increase more than proportionally than the promotive effects exerted by the red component, showing that lettuce-seed germination can be inhibited with a R:FR ratio close to one if they were exposed to high irradiances. This would result in inhibition of germination with high R:FR values through the high irradiance response system (HIR), when seeds are subjected to prolonged exposures to light of high PFD. Similarly, Hendricks et al. (1968) showed that the highest inhibitory effect on Amaranthus arenicola seeds was exerted with prolonged exposures to R light subtly enriched with FR wavelength bands around 720 nm (probably a similar light environment as the one prevailing below an incipient canopy). These observations, together with the results reported in experiment no. 2, where only prolonged exposure to canopy-filtered light did inhibit germination, suggest that a HIR mediated response is involved in the regulation observed in this work. As suggested by the much lower germination percentages attained in bare-soil conditions than in the laboratory after a saturating R pulse, it appears that both Ch. album and A. quitensis did not require such a FR enrichment but could be inhibited by prolonged exposure to direct sunlight (Fig. 4). This would make them unable to detect the presence of an establishing canopy.
Although the reasons why the species studied showed two different types of response to light quality (seeds that can or cannot detect the presence of an establishing canopy) are largely unknown, differences between seed-tegument optical properties cannot be ruled out. For example, Hendricks et al. (1968) found that A. arenicola seed germination could not be inhibited by continuous irradiation with light of 400–500 nm, because of the low transmittance of the seed coat in that region.
These results are the first to show that in some species (B. campestris and S. gallica), direct solar radiation subtly enriched with FR can exert an inhibitory effect on germination under field conditions. This capacity to detect small environmental light-quality modifications when seeds are exposed to high irradiances, allows them to detect the presence of a canopy in very early stages of its establishment and to remain in ‘safe’ pre-germination phases when the probability of successful seedling establishment is low. In the case of S. gallica, which has a seed diameter of less than 1 mm and seeds which do not germinate in the dark, seedlings would probably not reach reproductive stages if they were to emerge simultaneously with the crop, owing to a poor competitive ability in this developmental stage.
Further work on the ecological significance under field conditions of previously light responses described in the laboratory, will probably reveal that plants are able to acquire much more information from the light environment than we currently suspect.
We thank Dr R. A. Sánchez and Dr J. J. Casal for useful comments on the manuscript, and V. Passarella for providing technical assistance. This research was financially supported by Fundación Antorchas (project A-13359/1–000046) and SECYT-CONICET (BID 802/OC-AR PID N°1/344). D. Batilla is recipient of a CONICET Fellowship.