Variation within taxon
On several occasions there were larger differences in cumulative germination within taxa rather than between them (Figs 1 and 3). Part of the intra-taxon variation depended on the collection date (Tables 3 and 4) or starting point in the cycles (Table 4), but there were also substantial differences between seed batches collected on the same day and subjected to the same treatment (Figs 1 and 3). Especially large differences within a taxon were found in P. dubium ssp. lecoqii: of four seed batches only one germinated to a substantial extent at any of the circumstances tested (Figs 1 and 3). Pronounced differences occurred also within P. dubium ssp. dubium (Fig. 3), even though the two sites used for this taxon were the two most adjacent sites used (Table 1).
The differences within each taxon, when being subjected to the same treatment (Figs 1 and 3), are regarded as a result of different dormancy strengths. Compared with a seed batch with weak dormancy, a seed batch with strong dormancy requires longer and/or closer to optimum and/or repeated environmental event(s) to reduce the degree of dormancy sufficiently to reach a specific germination fraction in any specific environment. It seems possible that large differences in dormancy strength within Papaver taxa are not rare. Grime et al. (1981) tested one seed batch of each of P. argemone, P. rhoeas and P. dubium: for P. argemone, no suitable germination environment was found, but for both P. rhoeas and P. dubium, germination occurred to a substantial extent at 5 °C in darkness, in contrast to results from other Papaver studies (Fig. 2; Milberg and Andersson, 1997; Baskin et al., 2002). However, the study by Grime et al. (1981) did not aim to investigate dormancy, and seeds had been stored dry for some time before being tested; a treatment that we have observed, per se, reduces dormancy in Papaver, even though we have not seen such extensive germination (data not shown). Substantial differences of 10–40 % within a year (Baskin et al., 2002) were also found among P. rhoeas seed batches when collected at different sites within 15 km on the same day and sown outdoors at one single site.
Despite the variation within each taxon, taxon was an important exploratory factor for germination in annual cycles (Table 4). In addition, when subjected to a continuous temperature regime there were taxon-specific temperature preferences (Table 3). Therefore, it was relevant to investigate the differences in germination timing between taxa on a general level, instead of focusing on the specific fraction of germination at any single moment. It is also obvious that comparative studies need to include several seed batches per taxon studied to be able to reach sound conclusions regarding inter-taxa differences. Complementary to such extensive studies, some kind of general description of dormancy pattern and germination preferences may be used in order to evaluate potential differences between taxa or locally adapted types.
Seed dormancy classification
Papaver has relatively small embryos in comparison with the seeds (Martin, 1946), the embryos have to grow before germination and Papaver seeds imbibe easily. In accordance with this, Papaver has, according to the classification system proposed by Baskin and Baskin (2004), either (1) morphological dormancy (MD) or (2) some kind of morphophysiological dormancy (MPD), which means that (1) embryo growth and germination occur as one continuum within 4 weeks when seeds are placed in an environment suitable for germination or that (2) there is a (hypothetical) physiological inhibiting mechanism that has to be disarmed before germination can occur during any circumstances.
Seed dormancy of P. rhoeas is reduced when incubated in darkness at moderate temperatures (approx. 20–25 °C), and embryo growth and germination occur over a wide range of temperatures when provided with light during daytime (Milberg and Andersson, 1997; Baskin et al., 2002). Therefore, as concluded by Baskin et al. (2002), P. rhoeas has non-deep simple MPD.
The present results provides evidence that the other three Papaver taxa studied also belong to the non-deep simple MPD dormancy described by Baskin and Baskin (2004), even though the experimental set-up in this study aimed to investigate germination timing, not to classify dormancy. First, because no environment led to substantial germination for fresh seeds (Fig. 1), MD is excluded. Second, fresh seed of all taxa responded positively to a reduction in temperature, during the first autumn, in the annual cycles (Fig. 3). Third, the non-deep simple MPD dormancy is the only type of MPD in the scheme that allows only one warm period for dormancy reduction (the other types have requirements of either warm plus cold, cold plus warm plus cold, or only cold). Thus, if the Papaver taxa should belong to any other kind of MPD, the first sign of reduced dormancy should have appeared at the earliest during or after the first winter (Fig. 3). However, full germination, which is needed to regard a treatment as successful in completely breaking dormancy (sensu Baskin and Baskin, 2004), was not achieved in this study. Therefore, the possibility of an as yet undescribed type of dormancy exists. However, an ongoing study indicates that environmental conditions suitable for dormancy reduction of P. argemone and the two P. dubium subspecies are similar to those for P. rhoeas, but requiring longer periods (data not shown). Thus, we conclude that seed dormancy classification should be non-deep simple MPD (Baskin and Baskin, 2004) for all four taxa studied. Therefore, neither the dormancy classification scheme of Baskin and Baskin (2004) nor other classification systems proposed (Harper, 1957, 1977; Lang, 1987) can be used to separate differences in germination timing for the taxa studied. Furthermore, the non-deep simple MPD classification covers both species that reduce dormancy during a warm period and those that reduce dormancy during a cold period (Baskin and Baskin, 2004), and is therefore not informative when aiming to understand, compare or predict germination in an ecological perspective.
To facilitate comparisons, and also discussion, it is practical to structure the description of the process leading to germination. Assuming that seed batches are the study objects, it may be possible to describe species/populations by: (1) ‘dormancy pattern’, i.e. what are the environmental events that reduce and, if applicable, induce dormancy; (2) ‘germination preferences’, i.e. what environments are (or became during dormancy reduction) suitable for germination; and (3) ‘dormancy strength’, i.e. how much effort is needed to reduce dormancy.
Germination timing
Climate was a strong explanatory factor for germination (Table 4). The colder the climate, the more germination occurred in spring instead of autumn and the less germination occurred during the last year of the study (Figs 3 and 5). Germination in spring may be explained either by dormancy being reduced under cold conditions (0 °C), or by the seeds being on the way to germinate before winter, but the actual germination being postponed until suitable temperatures, or sufficient temperature in total, occurred in spring. In the cold climate, most germination occurred in the first autumn, but later in the study nearly all germination occurred during spring, and almost exclusively so for P. rhoeas (Fig. 3). Either P. rhoeas was the only taxon for which 20/10 °C (summer in the cold climate) was high enough for dormancy reduction, or the other taxa required longer times than 90 d for dormancy reduction at that temperature. Given that P. rhoeas is known to reduce dormancy within a time period of 12 weeks, albeit in darkness, at 20/10 °C, partly at 15/5 °C but not at 1 °C (Baskin et al., 2002), the explanation that germination during spring is a result of germination being delayed during cold periods seems to be the most plausible.
Several species are reported as having dormancy cycles (e.g. Baskin and Baskin, 1985; Murdoch, 1998; Mennan and Nguajio, 2006), i.e. after dormancy reduction, induction of dormancy occurs, as a response to an environmental event, if circumstances acceptable for germination have not been present. Among species with reported dormancy cycles is P. rhoeas (Milberg and Andersson, 1997; Cirujeda et al., 2006), in which dormancy is induced during cool periods, thus avoiding spring germination. In the intermediate and cold climates, germination occurred in autumn and in spring directly after transference from 0 to 5 °C (Fig. 3). Thus, seeds remaining ungerminated after autumn had a degree of dormancy low enough to allow at least some germination after the winter. This may be the result of 0 °C and –12 °C being too low for induction of dormancy, or of only –12 °C being too low and an ongoing germination process continuing at 0 °C, but too slowly for germination during the winter period.
There was a difference between times of collection (Tables 3 and 4), with the late-collected seeds having weaker dormancy and therefore germinating more easily (Figs 1 and 3). Some species exhibit differences in germination characteristics depending on whether seeds are from plants that emerged during spring or autumn. For Capsella bursa-pastoris in Sweden (Baskin et al., 2004) there was a difference observed when seeds were subjected to suboptimal conditions. Most germination occurred for seeds from mother plants that emerged in spring, but the difference was neutralized when seeds were subjected to optimal conditions or after a winter outdoors. For Galium aparine and Brassica kaber (syn. Sinapis arvensis) in Turkey (Mennan and Nguajio, 2006) there was a distinct difference in time for germination, with the seeds from spring mother plants mainly germinating earlier than those from autumn mother plants, even 2 years after collection. The emergence date for the plants used for our Papaver seed collections cannot be confirmed, but from field observations it is likely that the July (summer) and the remaining (autumn) collections were from mother plants that emerged during autumn and spring, respectively. For Papaver, the difference between summer- and autumn-collected seeds was only pronounced for initial germination; in treatments where additional germination occurred, the differences tended to level out with time (Fig. 3).
Regardless of collection time and starting point in cycles, time for germination, or emergence, was similar for all batches within a taxon (Figs 3 and 5). Therefore, differences in dormancy strength due to collection date for the four Papaver taxa studied here are probably explained by response to ripening circumstances and not an adaptation leading to two distinctly different generations. A stronger dormancy when ripening during a warmer period (summer) than during a cooler period (autumn) seems to be a proper adaptation to avoid germination during occasional periods with circumstances suitable for germination during summer, which would be important not least in the Mediterranean region, where these taxa probably evolved (Kadereit, 1990). This contrasts with Avena fatua for which a higher temperature during maturation reduces dormancy (Peters, 1982).
Comparison of taxa
Germination preferences differed between taxa. Compared with P. rhoeas and P. dubium ssp. dubium, P. argemone preferred lower temperatures for germination (Figs 1 and 3). When subjected to 30/20 °C for 900 d, full germination occurred for P. rhoeas and P. dubium ssp. dubium, and for one seed batch of P. dubium ssp. lecoqii, but there was no germination of P. argemone (Fig. 1), even though 30/20 °C as summer temperature was suitable for dormancy reduction for all taxa (Fig. 3). Mortality occurred exclusively for P. argemone subjected to 30/20 or 25/15 °C in light for more than 480 d. The mortality may be an additional indicator that P. argemone requires lower temperatures for germination than for the others tested; for the other taxa, the viability of the remaining seeds was nearly 100 % after 900 d in light at 30/20 or 25/15 °C. The relative similarity in germination temperature preferences between P. rhoeas and P. dubium ssp. dubium compared with P. argemone is reflected in the close relationship between P. rhoeas and P. dubium ssp. dubium (Kadereit et al., 1997; Carolan et al., 2006).
The significant interaction taxon x collection x start (Table 4) was partly a result of P. argemone germinating more after autumn than after a summer start for autumn-collected seeds in the cold climate (Fig. 3), while the other taxa did not differ with starting point in the cold climate, and all taxa, including P. argemone, germinated more after a summer start than after an autumn start in the intermediate and warm climates (Fig. 3). This pattern can be explained by a summer start at intermediate and warm climates including a period warm and long enough to lower the degree of dormancy and to lead to germination during autumn, and that the autumn start in the cold climate (5 °C) was well suited for germination for fresh seeds of P. argemone (Fig. 1). Because of stronger dormancy for the summer-collected seeds, the germination response directly after start in the cold climate occurred mostly for autumn-collected seeds (Fig. 3).
The differences in germination timing were mostly due to germination temperatures (Fig. 3), giving a later germination in autumn for P. argemone. However, germination in spring occurred simultaneously for all taxa (Fig. 3), showing that low temperatures did not restrict germination for P. rhoeas and P. dubium, as moderate temperatures did for P. argemone. Lack of germination at low temperatures in autumn for P. rhoeas and P. dubium was therefore probably a result of all possible germination occurring as soon as a suitable temperature was present, i.e. early during autumn for these two taxa.
Dormancy strength affected the overall response of the taxa. Even though P. rhoeas and P. dubium ssp. dubium germinated at similar temperatures (Fig. 4) and achieved nearly full germination after three autumns in the warm climate (Fig. 3), P. dubium ssp. dubium had its peak in germination during the second autumn whereas P. rhoeas germinated about equally in the first and the second autumns (Fig. 3). Thus, P. rhoeas had weaker dormancy than the other taxa. Other indications of P. rhoeas having the weakest dormancy are that this taxon was the only one that germinated during the second cycle in the cold climate (Fig. 3), most frequently germinated without a change of temperature condition (Fig. 1) and, to the highest degree, overcame the light preference for germination (Figs 2 and 4). Strongest dormancy was found in P. dubium ssp. lecoqii, which, despite the close relationship to P. dubium ssp. dubium, germinated very little (Fig. 3).
In an evolutionary perspective, it seems that the general dormancy pattern (dormancy reduction when warm) is a conservative character for the Papaver studied here. Germination temperature preferences have probably been adjusted when or after new taxa evolved, thus being less conservative than the general dormancy pattern. Dormancy strength, although to some degree being taxon-specific, varied between seed batches (Fig. 3), seemingly being the most easily changeable character involved in germination. One Papaver species, P. somniferum, has been reported to have weak dormancy without a requirement for a warm period before germination (Bare et al., 1978; Grime et al., 1981). It is a closer relative to P. rhoeas and P. dubium than to P. argemone (Kadereit et al., 1997; Carolan et al., 2006). These results may depend on the fact that P. somniferum ssp. somniferum, which has been cultivated since ancient times, has lost its dormancy because of artificial selection. Bare et al. (1978) described their study object as a ‘cultivated plant’, and only P. somniferum ssp. somniferum, the cultivated type, is known to occur outside the Mediterranean region in Europe (Tutin et al., 1964; Kadereit, 1986). Therefore, the population at Sheffield (Grime et al., 1981) was probably naturalized P. somniferum ssp. somniferum, not the wild-type P. somniferum ssp. setigetum.
Field implications
The four Papaver taxa studied here are annual weeds, occurring in crop fields. Therefore, a substantial part of the seeds from each cohort can be assumed to be buried. They are known to persist in soil for at least 5 years (Roberts and Boddrell, 1984), and P. dubium and P. rhoeas survived to 91 and 84 %, respectively, after 11 years in soil (Salzmann, 1954). Reduction of dormancy of Papaver being buried in soil will normally not lead to germination (Fig. 4). This is to be expected, because a light requirement for germination, not seed dormancy, is probably the most important factor for forming persistent seed banks (Thompson et al., 1997). Papaver rhoeas is known to show increased emergence in cultivated, compared with uncultivated, soil (Roberts and Feast, 1973) and it is possible to achieve 100 % germination of at least P. rhoeas using a combination of warm stratification in darkness and a germination test in cooler environments equipped with light (Milberg and Andersson, 1997; Baskin et al., 2002). However, in the field, the entire cohort of any of the Papaver taxa studied here will not germinate during a single season (Figs 3 and 5), even if not buried in soil and thus prevented from germination by lack of light (Milberg and Andersson, 1997). Instead, the seed dormancy pattern in combination with germination requirements and dormancy strength distributes the emergence of a cohort over several seasons, regardless of local climate, weather and soil cultivation methods. From a weed management point of view, P. rhoeas is a problematic taxon (Holm et al., 1997) that is frequently reported to show herbicide resistance (e.g. Paterson et al., 2002; Durán-Prado et al., 2004; Scarabel et al., 2004). For the reasons described above, the old method of fallow practice, i.e. allowing germination but not flowering, would probably be the best to reduce the seed bank in heavily infested fields.
Because of the extensive variation within each of the taxa studied (Fig. 3), despite being collected within a relatively small area (Table 1), taxon-wise predictions about germination, as fractions, are not possible. By contrast, the general response to climate and seasonal changes was distinct and taxon-specific (Fig. 3). Therefore, it should be possible to predict the relationship between autumn- and spring-germinated parts, and germination timing for seeds not buried in the soil, with relatively good certainty, from temperatures during the last year before the specific occasion, at least in climates where seeds remain imbibed for most of the year (cf. Fig. 3 and Fig. 5). For example, a warm summer would have resulted in extensive germination during autumn (because of the wide range of temperatures accepted for germination) and thus little germination would have been possible the following spring. However, if soil disturbance transfers seeds to the soil surface, germination can occur in a seemingly unpredictable pattern (Roberts and Boddrell, 1984) as a result of seed being subjected to suitable dormancy reduction circumstances but remaining ungerminated in soil until they are exposed to light (Milberg and Andersson, 1997).
As shown in this study, these Papaver taxa perform as winter annuals in warmer climates while spring emergence is an important factor in cooler climates (Figs 3 and 5). Therefore, changes in emergence pattern as a result of climate changes are also predictable: a warmer climate will decrease spring emergence, and increase autumn emergence, in relation to the present situation.
Answers to all your Biology Questions
