Introduction
Physical dormancy (PY) in seeds, which is caused by a water-impermeable seed/fruit coat, has long been considered to be a mechanism ensuring that germination occurs at a favourable time for seedling growth and establishment (Rolston, Reference Rolston1978; Baskin, Reference Baskin2003; Baskin and Baskin, Reference Baskin and Baskin2014a). In contrast, Paulsen et al. (Reference Paulsen, Colville, Kranner, Daws, Hogstedt, Vandvik and Thompson2013) proposed the hypothesis that PY evolved as a predator escape mechanism, especially to escape from seed-eating rodents, and they (Paulsen et al., Reference Paulsen, Högstedt, Thompson, Vadvik and Eliassen2014) further developed this argument by testing this phenomenon with a mathematical model. As Paulsen et al. (Reference Paulsen, Colville, Kranner, Daws, Hogstedt, Vandvik and Thompson2013) suggested, this new interpretation of the evolution of PY has renewed interest in the function and evolutionary history of this class of seed dormancy. However, the question whether PY evolved as a predator-escaping mechanism needs to be considered within the broad framework of what is known about its ecological adaptive value. We believe that some misconceptions about the evolution of PY have been incorporated into this hypothesis, and it does not consider other important selective pressures. Therefore, the objective of this opinion paper is to critically evaluate and discuss the validity of the crypsis hypothesis (Paulsen et al., Reference Paulsen, Colville, Kranner, Daws, Hogstedt, Vandvik and Thompson2013, Reference Paulsen, Högstedt, Thompson, Vadvik and Eliassen2014) on the evolution of PY.
Physical dormancy in seeds
Water impermeability of seeds or fruits has been recognized as a mechanism of seed dormancy (i.e. PY) in various seed dormancy classification systems (Crocker, Reference Crocker1916; Nikolaeva, Reference Nikolaeva1969; Harper, Reference Harper1957, Reference Harper1977; Baskin and Baskin, Reference Baskin and Baskin2004, Reference Baskin and Baskin2014a; Hilhorst, Reference Hilhorst and Kermode2011). In their hierarchical scheme of seed dormancy classification, i.e. classes, levels and types, Baskin and Baskin (Reference Baskin and Baskin2004) considered PY to be one of five classes of dormancy, but they did not include any levels or types in the class. Embryos of most of the seeds/fruits (diaspores) with PY are non-dormant, and thus the seeds can germinate in a wide range of environmental conditions after PY is broken (Baskin and Baskin, Reference Baskin and Baskin2014a). However, some species produce diaspores with combinational dormancy, i.e. the embryo is physiologically dormant, and diaspores have a water-impermeable coat, e.g. Ceanothus spp. Cercis spp., Cotinus spp., Diplopeltis huegelii, Geranium spp., Rhus spp. and Tilia spp. (Baskin and Baskin, Reference Baskin and Baskin2014a). Thus, in these species both PY and PD have to be broken prior to germination.
Although there are many claims that PY can be broken via abrasion by soil particles or microbial activity, there is little or no experimental evidence to support these claims (Baskin and Baskin, Reference Baskin and Baskin2000). However, it has been shown that temperature- and moisture-dependent environmental cues break PY in seeds of many species (Baskin, Reference Baskin2003; Baskin and Baskin, Reference Baskin and Baskin2008, Reference Baskin and Baskin2014a). When a physically dormant seed/fruit receives a dormancy breaking environmental cue, i.e. fluctuating temperatures (Taylor, Reference Taylor1981, Van Assche et al., Reference Van Assche, Debucquoy and Rommens2003), high temperatures (Keeley, Reference Keeley1987; Herranz et al., Reference Herranz, Ferrandis and Martinez-Sanchez1998, Morrison et al., Reference Morrison, McClay, Porter and Rish1998), low temperatures (Gama-Arachchige et al., Reference Gama-Arachchige, Baskin, Geneve and Baskin2012), moist heat (Jayasuriya et al., Reference Jayasuriya, Baskin and Baskin2008a, Reference Jayasuriya, Baskin, Geneve and Baskinb) or dry heat (Jayasuriya et al., Reference Jayasuriya, Baskin and Baskin2009a), an opening is formed in a specialized morpho-anatomical area in the seed/fruit coat (Baskin et al., Reference Baskin, Baskin and Li2000). This specialized morpho-anatomical area is termed the ‘water gap’, and Gama-Arachchige et al. (Reference Gama-Arachchige, Baskin, Geneve and Baskin2013) have identified 24 different kinds of ‘water gaps’ in 16 families of angiosperms containing seeds with PY. To date, 18 angiosperm families belonging to five major plant clades (magnoliids, monocots, core-eudicots, rosids and asterids) have been identified as producing seeds or fruits with PY (Mahadevan and Jayasuriya, Reference Mahadevan and Jayasuriya2013). Thus, PY of seeds has evolved independently several times during the evolution of angiosperms (Finch-Savage and Leubner-Metzger, Reference Finch-Savage and Leubner-Metzger2006; Baskin and Baskin, Reference Baskin and Baskin2014a).
Since formation of the opening in the ‘water gap’ during PY break is a physical process, resealing of the opening is not possible (Jayasuriya et al., Reference Jayasuriya, Baskin, Geneve and Baskin2009b). Thus, once PY is broken it cannot revert back to dormancy, i.e. cycling of dormancy ↔ non-dormancy is not possible in seeds with PY. However, seasonal cyclic patterns of seed germination have been reported for some species producing seeds with PY (Taylor and Revell, Reference Taylor and Revell1999; Van Assche et al., Reference Van Assche, Debucquoy and Rommens2003; Benvenuti et al., Reference Benvenuti, Dinelli, Bonetti and Catizone2005; Jayasuriya et al., Reference Jayasuriya, Baskin and Baskin2008a, Reference Jayasuriya, Baskin, Geneve and Baskinb). Jayasuriya et al. (Reference Jayasuriya, Baskin and Baskin2008a, Reference Jayasuriya, Baskin, Geneve and Baskinb, Reference Jayasuriya, Baskin, Geneve and Baskin2009b) described this seasonal cyclic pattern of germination in seeds with PY using a sensitivity cycling model. According to the model, seeds cycle between sensitive ↔ insensitive states to dormancy breaking treatment, but seeds do not cycle between dormancy ↔ non-dormancy (Jayasuriya et al., Reference Jayasuriya, Baskin and Baskin2009a).
Evolution through natural selection versus crypsis hypothesis and ‘evolution’ of PY
Although several models and theories have been put forward to explain evolution at the population and genetic levels (Wright, Reference Wright1932; Dawkins, Reference Dawkins1976: Maynard Smith, Reference Maynard Smith1978, Reference Maynard Smith1982: Eshel, Reference Eshel1983; Hammerstein and Selten, Reference Hammerstein, Selten, Aumann and Hart1994; Hammerstein, Reference Hammerstein1996), the basic theory of evolution through natural selection of C. Darwin and A.R. Wallace has not changed since its proposal in 1859. According to Darwin (Reference Darwin1859) and Wallace (Reference Wallace1859), variations among individuals in a population are the raw materials for evolution, and natural selection is the mechanism of evolution. Further, variation among individuals occurs randomly, and different selection pressures in the environment select the character state best fitted for the existing environment. Random variation between individuals in a population is the key point in the evolutionary theory of Darwin (Reference Darwin1859) and differs from the evolutionary theory of Larmarck, who explained variation of an acquired feature of the organism as an adaptation to the prevailing environmental conditions (Gillispie, Reference Gillispie1958).
In the crypsis hypothesis, Paulsen et al. (Reference Paulsen, Colville, Kranner, Daws, Hogstedt, Vandvik and Thompson2013, Reference Paulsen, Högstedt, Thompson, Vadvik and Eliassen2014) claim that PY evolved to allow seeds to escape from predators, especially rodents. However, this claim is incomplete with regard to Darwin's explanation of evolution that variations occur randomly within populations. If we assume that PY has evolved from a physiologically dormant (PD) seed-producing ancestor (see Willis et al., Reference Willis, Baskin, Baskin, Auld, Venable, Cavender-Bares, Donohue and de Casas2014), PY is the variation that occurred in the population of seeds with PD. Thus, according to Darwin's explanation, it is just a random variation within the population. Therefore, PY did not have a function when it evolved initially, unless Larmarck's hypothesis on evolution or creationism is considered to be true. As such, then, the main argument of the crypsis hypothesis of evolution of PY cannot be accepted. Willis et al. (Reference Willis, Baskin, Baskin, Auld, Venable, Cavender-Bares, Donohue and de Casas2014) hypothesized that PY may have evolved from an ancestor that produced PD seeds. Thus, PY is a variation in the plant population with individuals producing seeds with PD. This variation arose randomly in the population as a mutation or a series of mutations. Several selection pressures may have been operating on the population, e.g. predator pressure, pathogen pressure and pressure in timing germination to occur at the best time for seedling survival and establishment. Thus, not only escape from predators but also other ecological circumstances may have contributed to the fitness of the individuals via natural selection for PY, and thus it was selected to be retained in the environment.
Other shortcomings of the crypsis hypothesis
Paulsen et al. (Reference Paulsen, Colville, Kranner, Daws, Hogstedt, Vandvik and Thompson2013) stated that seeds with PY cannot do anything better than seeds with PD can do when timing of germination is considered. Thus, they further raised the question of why PY evolved. However, it seems that this is a general phenomenon that can be seen throughout evolution, i.e. there are several examples of more than one strategy performing the same function. Thus, just because seeds with PY cannot do anything better than those with PD does not invalidate the fact that PY plays a significant role in timing of germination to the correct period for seedlings to establish, i.e. it is another way of achieving the same outcome. Water-impermeable coats in sexually regenerating structures are not restricted to angiosperms. Even though water-impermeable seed coats are absent in regenerating structures of gymnosperms, the aquatic pteridophyte Marsilea has a sporocarp with water-impermeable walls and even a water-gap structure. Unlike the water-impermeable coats of angiosperm seeds or fruits, which enclose the embryo, the water-impermeable coats of sporocarps enclose microspores and megaspores that give rise to the gametophytic generation. However, both structures are involved in regulating reproduction (Gama-Arachchige, Reference Gama-Arachchige2013). Thus, water-impermeable coats have been involved in regulation of regeneration time, perhaps before the evolution of angiosperms.
Furthermore, there are some adaptive advantages of PY over PD. PY break is a physical process and thus occurs much faster than PD break, which is a physiological process. Dormancy breaking of seeds with PY occurs before they are imbibed, whereas those with PD may need to be imbibed for a few months while dormancy break occurs. For example, sensitive seeds of Ipomoea lacunosa (Convolulaceae), which have PY, require only 3 h of exposure to 35°C moist conditions for dormancy to be broken (Jayasuriya et al., Reference Jayasuriya, Baskin and Baskin2008a). Further, after PY is broken most seeds germinate within 12–48 h. Thus, PY functions as a barrier to water entrance into the seed until a short time prior to germination (Baskin and Baskin, Reference Baskin and Baskin2004).
Paulsen et al. (Reference Paulsen, Colville, Kranner, Daws, Hogstedt, Vandvik and Thompson2013) further suggested that production of volatiles by imbibed seeds ‘… could explain why hard seeds are surprisingly frequent in hot deserts…’. That is, PY has ‘arisen in plants of (at least seasonally) dry environments’ to deal with the problem of seed predation. Baskin and Baskin Reference Baskin and Baskin(2014a) showed that PY occurs in all the biomes on Earth, with the highest frequency in tropical deciduous forests, tropical savannas and hot deserts, with a second peak in temperate steppes, matorral and cold deserts; however, in all these arid vegetation regions PD is more common than PY. Meisert (Reference Meisert2002) divided the germination strategies of 35 Geraniaceae species into three categories: PY100, seed lots with >80% PY seeds; PY80, seed lots with < 80% PY seeds; PY0, non-dormant seeds (100% were water permeable). Further, according to Meisert PY0 species are succulents that inhabit arid environments. Thus, within the PY-producing family Geraniaceae it seems that there are more non-dormant than PY species inhabiting arid desert ecosystems. If, as suggested by Paulsen et al. (Reference Paulsen, Colville, Kranner, Daws, Hogstedt, Vandvik and Thompson2013), seed-eating predators are the only selection pressure operating on the kind of seed dormancy, PY should be dominant among the arid desert species of Geraniaceae. Further, PY in arid environments may also be an adaptation to prevent ultra-drying of seeds. Ultra-drying of seeds [3–4% moisture content (MC)] can increase the formation of antioxidants through autoxidation of lipids in the seeds (McDonald, Reference McDonald1999). Thus, ultra-drying of seeds can increase seed deterioration. The water-impermeable palisade layer may play a role in preventing seeds with PY from ultra-drying.
Paulsen et al. (Reference Paulsen, Högstedt, Thompson, Vadvik and Eliassen2014) proposed an argument that is contradictory to that of Paulsen et al. (Reference Paulsen, Colville, Kranner, Daws, Hogstedt, Vandvik and Thompson2013) regarding the high occurrence of PY seeds in dry environmental conditions. According to the model proposed by Paulsen et al. (Reference Paulsen, Högstedt, Thompson, Vadvik and Eliassen2014), the probability of occurrence of soft seeds is proportional to the probability of rodents detecting them under prevailing environmental conditions. The probability of detecting soft seeds by a rodent is proportional to the amount of volatiles released by the seeds, which is proportional to the environmental humidity (Paulsen et al., Reference Paulsen, Högstedt, Thompson, Vadvik and Eliassen2014). Under dry environmental conditions, a low amount of volatiles is released even by the soft seeds. Thus, probability of detecting hard seeds is equal to probability of detecting soft seeds under dry environmental conditions. As such, then, there is no point in investing in a hard seed coat if hard seededness is significant only in ‘predator escape’ in arid ecosystems.
According to Paulsen et al. (Reference Paulsen, Colville, Kranner, Daws, Hogstedt, Vandvik and Thompson2013, Reference Paulsen, Högstedt, Thompson, Vadvik and Eliassen2014), the only way to explain the presence of impermeable and permeable (‘dimorphic’) seeds in the same seed cohort is that water-permeable seeds are eaten by rodents, but the impermeable ones are taken (but not eaten) and cached. Since the water-impermeable seeds do not produce volatile compounds, the rodents may not be able to find the seeds they cached, in which case these seeds may become part of a persistent seed bank and thus lower the effects of reproductive failure in years when no seeds are produced. Paulsen et al. (Reference Paulsen, Colville, Kranner, Daws, Hogstedt, Vandvik and Thompson2013) suggested that the permeable seeds are a reward for rodents for their dispersal of dormant seeds. However, even if we accept the hypothesis suggested by Paulsen et al. (Reference Paulsen, Colville, Kranner, Daws, Hogstedt, Vandvik and Thompson2013), both hard seeds and soft seeds should be consumed similarly as soon as they are detected. If soft seeds are detected in higher numbers, then they will be consumed in higher numbers, and thus the selective pressure would have already eliminated the soft seeds. Paulsen et al. (Reference Paulsen, Högstedt, Thompson, Vadvik and Eliassen2014) have considered the rate of soft-seed detection and consumption in different environmental conditions as a parameter of determining the proportion of soft seeds produced by a species. However, during model preparation they overlooked the fact that hard seeds have to be softened prior to germination. Thus, during the time of dormancy break to germination they have the same risk of being detected by a predator as do soft seeds during the time from dispersal to germination. The amount of time required for soft seeds to germinate would be similar to the time required for dormancy break and germination of hard seeds.
Paulsen et al. (Reference Paulsen, Colville, Kranner, Daws, Hogstedt, Vandvik and Thompson2013) stated that PY is absent in small-seeded lineages among angiosperms. Although they cite Leishman and Westoby (Reference Leishman and Westoby1998), Sera and Sery (Reference Sera and Sery2004) and Moles et al. (Reference Moles, Ackerly, Webb, Tweddle, Dickie and Westoby2005) to support this argument, none of these studies conclude directly that PY is absent in small-seeded lineages in angiosperms. The major conclusion of Leishman and Westoby (Reference Leishman and Westoby1998) is that seed mass is not associated with soil seed-bank persistence. Sera and Sery (Reference Sera and Sery2004) concluded that the variation in seed size in their dataset correlates mainly with the plant's reproductive capacity, life cycle, ability to propagate vegetatively and seed dispersal ability. Moles et al. (Reference Moles, Ackerly, Webb, Tweddle, Dickie and Westoby2005) concluded that divergence in seed size is often associated with divergence of growth form of the plant. Thus, these studies provide no direct evidence that PY is absent in small-seeded lineages.
It is not clear how Paulsen et al. (Reference Paulsen, Colville, Kranner, Daws, Hogstedt, Vandvik and Thompson2013) define ‘small seeds’, and we have to ask if they consider seeds of legumes (with PY), such as those of Trifolium campestre with a mass of 0.0004 ± 0.0001 g (Baskin and Baskin, unpublished data), to be small or large. Further, for 100 tropical Fabaceae species, small to moderately sized seeds (0.002–0.075 g) had PY and large seeds had non-dormancy (0.06–20.7 g) or physiological dormancy (1.03–4.99 g) (Jayasuriya et al., Reference Jayasuriya, Wijetunga, Baskin and Baskin2013). About 50% of the species with small seeds ( < 0.05 g) had PY. Thus, the argument of Paulsen et al. (Reference Paulsen, Colville, Kranner, Daws, Hogstedt, Vandvik and Thompson2013) that no small seeds have PY does not seem to be valid. Further, even if PY is predominantly absent among small-seeded lineages, there exist several alternative explanations, such as small seeds are produced in large numbers following the opportunistic strategy (r) or that surface per volume ratio (seed coat and seed mass, respectively) in smaller volumes is higher, which makes the investment in a hard coat less economical in smaller seeds. Paulsen et al. (Reference Paulsen, Högstedt, Thompson, Vadvik and Eliassen2014) refer to a question raised by Baskin and Baskin (Reference Baskin and Baskin1998): ‘Why did physical dormancy evolve in so many families with recalcitrant seeds?’ Further, they (Paulsen et al., Reference Paulsen, Högstedt, Thompson, Vadvik and Eliassen2014) suggested that this may be because these are two alternative ways of escaping from predators: (1) recalcitrant seeds germinate fast and thus escape the risk of predation as soon as they are dispersed into the environment, which is favourable for germination and seedling establishment; and (2) on the other hand, PY seeds are in unfavourable conditions for germination during dispersal, and thus they need physical protection to escape from predators until the environment becomes favourable for seedling establishment. However, this argument is not 100% valid since: (1) in recalcitrant seeds of some species, there is a considerable time period between dispersal and germination (e.g. Yu et al., Reference Yu, Baskin, Baskin, Tang and Cao2008); and (2) many species with PY inhabit tropical rainforests, where conditions are favourable for seedling establishment at the time of seed dispersal (Jayasuriya et al., Reference Jayasuriya, Wijetunga, Baskin and Baskin2013; Baskin and Baskin, Reference Baskin and Baskin2014a).
Ecological significance of PY
In addition to escape from predators, there are other significant functions of PY that increase the fitness of the species, some of which will be discussed below.
Detection of suitable environmental conditions for seedling establishment
PY in seeds has long been recognized as a mechanism for detection of environmental conditions favourable for seedling establishment, such as post-winter and post-summer conditions, canopy gaps and post-fire environments (Rolston, Reference Rolston1978; Baskin, Reference Baskin2003; Baskin and Baskin, Reference Baskin and Baskin2014a).
Winter/summer detection
Several examples of the significance of PY in detecting environmental conditions favourable for seedling establishment are from the temperate zone, where the seasons are better demarcated than they are in the tropical zone. Summer annuals are adapted for growth in spring–summer and winter annuals for growth in winter–early spring; thus, seeds of summer annuals germinate in spring and winter annuals in autumn. Seed PY contributes significantly to ensuring that germination occurs at the correct time for seedling survival and plant establishment. Some of the many examples of this timing are included in research findings of Quinlivan (Reference Quinlivan1961), Cardina and Sparrow (Reference Cardina and Sparrow1997), Van Assche and Vandelook (Reference Van Assche and Vandelook2006), Ooi et al. (Reference Ooi, Auld and Denham2009) and Gama-Arachchige et al. (Reference Gama-Arachchige, Baskin, Geneve and Baskin2012). The significance of PY in detecting summer conditions will be discussed using Ipomoea lacunosa as an example.
Jayasuriya et al. Reference Jayasuriya, Baskin and Baskin(2008a) have clearly shown the significance of PY in I. lacunosa seeds in the timing of germination to a favourable time for seedling establishment. I. lacunosa is a summer annual weedy vine that germinates in early summer in soybean and corn fields in eastern USA. This species produces and disperses seeds in early to late autumn. Any seeds that are non-dormant germinate soon after dispersal in autumn, since environmental conditions at this time are favourable for them to do so. If seeds germinate in autumn, however, none of the seedlings survive the harsh winter conditions; they are killed by their first exposure to freezing temperatures. However, most I. lacunosa seeds have PY and remain in the soil seed bank during winter. Seed dormancy break in I. lacunosa can be explained with a two-step model (Jayasuriya et al., Reference Jayasuriya, Baskin and Baskin2008a). At the time of dispersal, most seeds are insensitive to dormancy breaking cues. Insensitive seeds require moderate temperature (~25°C)–moist conditions to become sensitive, and then sensitive seeds require a short period (2–4 h) of exposure to high temperature (~35°C)–moist conditions to become non-dormant. Thus, seeds become sensitive during spring and can become non-dormant during summer, when conditions are favourable for seed establishment and growth of plants to maturity.
Gap detection
Gaps are openings created in the existing vegetation due to removal of part of the canopy cover. In newly formed gaps, there is reduced competition and an increase in irradiance, red/far red (R/FR) ratio, temperatures, magnitude of daily temperature fluctuations and soil nutrients (e.g. nitrate), the combination of which is favourable for the growth of many high-light requiring species. Further, environmental signals in gaps may trigger dormancy break and/or germination of species with PY (Vazquez-Yanes and Orozco-Segovia, Reference Vazquez-Yanes and Orozco-Segovia1982, Reference Vazquez-Yanes, Orozco-Segovia, Caldwell and Pearcy1994). Fluctuating soil temperatures and soil moisture/high temperatures have been identified as the environmental signals that trigger the breaking of PY. Several researchers have reported higher percentages of germination of seeds of species with PY in gaps than under the closed canopy (Vazquez-Yanes and Orozco-Segovia, Reference Vazquez-Yanes and Orozco-Segovia1982, Reference Vazquez-Yanes, Orozco-Segovia, Caldwell and Pearcy1994; van Klinken et al., Reference van Klinken, Flack and Pettit2006). Further, convincing evidence that environmental cues in gaps trigger PY break has been collected in experiments in both field and laboratory conditions. Vazques-Yanes and Orozco-Segovia (Reference Vazquez-Yanes and Orozco-Segovia1982), De Paula et al. (Reference De Paula, Delgado, Paulilo and Santos2012) and Souza et al. (Reference Souza, Voltolini, Santos and Paulilo2012) identified simulated natural PY breaking treatments for Heliocarpus donnell-smithii (Malvaceae), Schizolobium parahyba (Fabaceae) and Cassia leptophylla (Fabaceae), respectively, that occur in forest gaps. For example, Vazquez-Yanes and Orozco-Segovia (Reference Vazquez-Yanes and Orozco-Segovia1982) showed that PY of seeds of the tropical pioneer species H. donnell-smithii can be broken by fluctuating temperatures (40/30 and 35/15°C) under laboratory conditions, and that these temperature fluctuations correspond to those in the centre and near the centre of a large forest gap in tropical rainforests in south-eastern Mexico, where dormancy break also occurs.
Post-fire detection
Fire-prone ecosystems are distributed throughout the world (Thonicke et al., Reference Thonicke, Venevsky, Sitch and Cramer2001), and fire is one of the major disturbance factors in many ecosystems (Thonicke et al., Reference Thonicke, Venevsky, Sitch and Cramer2001; Bond and Keeley, Reference Bond and Keeley2005). The post-fire environment is favourable for seedling establishment and plant development in fire-prone ecosystems for various reasons. Since fire totally or partially removes the existing vegetation, post-fire systems are free from intra- or interspecific competition. Further, burning the vegetation releases nutrients, thus fertilizing the soil and increasing nutrient mobilization. Removal of canopy shade provides favourable light conditions for seedlings of high-light requiring plants. Forest fires increase soil temperature to >100°C at shallow depths and, as a result, soils are sterilized and plant pathogens are reduced (Ahlgren and Ahlgren, Reference Ahlgren and Ahlgren1960; Keeley and Zedler, Reference Keeley and Zedler1978; Zedler, Reference Zedler, West, Shugart and Botkin1981; Thonicke et al., Reference Thonicke, Venevsky, Sitch and Cramer2001).
Much research has been conducted to determine the effect of fire on dormancy break, especially PY break (Baskin and Baskin, Reference Baskin and Baskin2014a). Ecologists working in fire-prone ecosystems have reported a sudden flush of seedlings after a fire, and prominent among the seedlings were species whose seeds have PY (Sweeney, Reference Sweeney1956; Hadley, Reference Hadley1961; Floyd, Reference Floyd1966; Shea et al., Reference Shea, McCormick and Portlock1979; Dell, Reference Dell1980; Abrams and Dickmann, Reference Abrams and Dickmann1982; Borchert, Reference Borchert1989; Moreira and Pausas, Reference Moreira and Pausas2012). There are various examples of dormancy break in PY seeds of obligate seeders by fire or by simulated fire temperatures (Floyd, Reference Floyd1966; Cushwa et al., Reference Cushwa, Martin and Miller1968; Keeley, Reference Keeley1987; Keeley and Keeley, Reference Keeley and Keeley1987; Gonzalez-Rabanal and Casal, Reference Gonzalez-Rabanal and Casal1995; Morrison et al., Reference Morrison, McClay, Porter and Rish1998; Williams et al., Reference Williams, Congdon, Grice and Clarke2003; Baeza and Roy, Reference Baeza and Roy2008; Bolin, Reference Bolin2009; Santana et al., Reference Santana, Bradstock and Ooi2010; Moreira and Pausas, Reference Moreira and Pausas2012). One such species is Iliamna corei, which is endemic to one site of Peters Mountain in Giles County, Virginia, USA (Hecht, Reference Hecht2008). PY in seeds of this species can be broken by fire (burning) and by dry (80°C for 30 min) or wet (boiling water for 1–20 s) heat. However, neither freezing and thawing, shifting from winter to summer temperature regimes nor long-term dry storage at ambient laboratory temperatures was successful in breaking PY (Baskin and Baskin, Reference Baskin and Baskin1997). A controlled burn at the population site on Peters Mountain stimulated hundreds of I. corei seeds to germinate (Hecht, Reference Hecht2008). Baskin and Baskin (Reference Baskin and Baskin1997) concluded that dormancy break in this species in nature is fire dependent.
Bradshaw et al. (Reference Bradshaw, Dixon, Hopper, Lambers and Turner2011) argued that the mass recruitment observed in several species that produce seeds with PY following wildfire can be viewed as an exaptation (pre-adaptation) rather than an adaptation to fire. They further argued that ‘… PY might have evolved from the key trait of persistence within the soil seed bank during unfavorable conditions’. Bradshaw et al. (Reference Bradshaw, Dixon, Hopper, Lambers and Turner2011) support their argument by citing that there are many species with PY in non-fire-prone ecosystems. However, Keeley et al. (Reference Keeley, Pausas, Rundel, Bond and Bradstock2011) argued that PY in non-fire-prone ecosystems might have evolved under selection from other stress factors, but that it may have been selected for in some species under the stress of fire in fire-prone ecosystems. Further, they cite Lupinus, Astragalus and Trifolium species (Fabaceae) as occurring in non-fire-prone ecosystems as well as in chaparral. Species in non-fire-prone ecosystems produce seeds with PY that is easily alleviated by high summer temperatures. However, species of the same genera that occur in chaparral produce seeds with PY that is broken by high fire temperatures. We favour the Keeley et al. (Reference Keeley, Pausas, Rundel, Bond and Bradstock2011) argument that in different ecosystems different stress factors may have been involved in selecting species producing seeds with PY.
Aid in seed dispersal
Although it has not been discussed extensively among seed scientists, there is a considerable amount of information in the literature on the significance of PY in seed dispersal. PY is significant mainly in two dispersal modes: hydrochory and endozoochory. In particular, there is a large amount of information in the literature on the significance of PY in endozoochory. Several studies have shown that a significantly higher percentage of non-scarified than of scarified seeds retained viability during passage through the guts of frugivores and large mammals (Glyphis et al., Reference Glyphis, Milton and Siegfried1981; Janzen et al., Reference Janzen, Demment and Robertson1985; Michael et al., Reference Michael, Steadman, Plummer and Vercoe2006). Further, other studies have shown that a considerable portion of seeds with PY retain viability during passage through the gut of horses, rodents, sheep, birds, lamas, donkeys and hares (Janzen, Reference Janzen1981; Miller, Reference Miller1995; Ramos et al., Reference Ramos, Robles and Castro2006; Campos et al., Reference Campos, Peco, Campos, Malo, Giannoni and Suarez2008). It seems that the water-impermeable seed coat is a barrier that prevents destruction of the embryo by digestive enzymes and digestive acids in the guts of frugivores and other herbivores (Janzen et al., Reference Janzen, Demment and Robertson1985; Michael et al., Reference Michael, Steadman, Plummer and Vercoe2006). Some studies found that germination percentages of intact seeds with PY were increased by passage through the guts of dispersal agents (Glyphis et al., Reference Glyphis, Milton and Siegfried1981; Izhaki et al., Reference Izhaki, Korine and Arad1995; Miller, Reference Miller1995; Ramos et al., Reference Ramos, Robles and Castro2006; Campos et al., Reference Campos, Peco, Campos, Malo, Giannoni and Suarez2008), whereas others reported that they were not (Janzen, Reference Janzen1981; Michael et al., Reference Michael, Steadman, Plummer and Vercoe2006).
Information on dispersal of seeds with PY through hydrochory is scant. Griffith and Forseth (Reference Griffith and Forseth2002) and Jayasuriya et al. (Reference Jayasuriya, Baskin, Baskin and Fernando2012) have speculated that PY in seeds of the hydrochorous Fabaceae species Aeschynomene virginica and Derris scandens, respectively, is an adaptation for water dispersal. During dispersal, seeds are exposed to conditions favourable for germination. Thus, if seeds were non-dormant they might germinate in water during dispersal, where conditions are not favourable for seedling establishment. Further, seeds with PY have an added advantage over those with PD or other classes of dormancy. The water-impermeable seed coat prevents seeds from imbibing during the period of water dispersal. Therefore, these seeds may remain viable longer than those with other kinds of dormancy. The water-impermeable testa not only acts as a barrier to water uptake but also as a barrier to salts (Nichols et al., Reference Nichols, Malik, Stockdale and Colmer2009). Guja et al. (Reference Guja, Merritt and Dixon2010) reported that seeds with PY have higher survival during dispersal in saline water than those with PD, but they stated that PY is not necessarily a requirement for seeds to be dispersed by saline water.
Escape mechanism from predators and pathogens
According to the crypsis hypothesis (Paulsen et al., Reference Paulsen, Colville, Kranner, Daws, Hogstedt, Vandvik and Thompson2013), the water-impermeable seed coat plays a major role in predator escape by preventing release of volatile compounds that allow would-be predators to locate the seeds. However, only a few data are available on this topic. Vander Wall (Reference Vander Wall1998), Jani and Jayasuriya (Reference Jani and Jayasuriya2013) and Paulsen et al. (Reference Paulsen, Colville, Kranner, Daws, Hogstedt, Vandvik and Thompson2013) have shown that when seeds with PY are scarified and then imbibed they produce more volatile compounds than dry non-scarified seeds. However, imbibed seeds that do not have PY also produce higher amounts of volatile compounds than non-imbibed ones (Vander Wall, Reference Vander Wall1998; Jani and Jayasuriya, Reference Jani and Jayasuriya2013). High metabolic rates, Maillard reaction and lipid peroxidation are some of the reasons why seeds produce more volatile compounds when they have a high moisture content (Colville et al., Reference Colville, Bradley, Lloyd, Pritchard, Castle and Kranner2012). Thus, predators can detect imbibed seeds more easily than non-imbibed ones (Johnson and Jorgensen, Reference Johnson and Jorgensen1981; Vander Wall, Reference Vander Wall1998; Jorgensen, Reference Jorgensen2001). Especially under moist conditions, seeds with PY have an advantage over those with other kinds of dormancy. That is, under moist conditions seeds without water-impermeable seed coats imbibe and produce more volatile compounds and will be detected by predators more easily than those with PY. Seeds with water-impermeable seed coats may not imbibe even though the soil moisture content is high. Thus, they would produce low, or no, amounts of volatile compounds and would not be detected by predators (Paulsen et al., Reference Paulsen, Colville, Kranner, Daws, Hogstedt, Vandvik and Thompson2013). Further, most seeds with PY germinate quickly after PY is broken, and thus they would escape predation.
Dalling et al. (Reference Dalling, Davis, Schutte and Arnold2011) have put forward a seed-defence theory, in which they suggest a close relationship between the seed dormancy class and the seed-defence syndrome. They identified three seed dormancy syndromes: physical defence syndrome, chemical and physical defence continuum and microbial defence syndrome. The physical defence syndrome is present in seeds with PY. Dalling et al. (Reference Dalling, Davis, Schutte and Arnold2011) suggested that the water-impermeable seed coat is a barrier to microbe access to the embryo. We further suggest that the water-impermeable seed coat prevents imbibition and thus keeps the moisture content low enough to be unfavourable for microbial action, and thus PY assists in reducing seed pathogenicity (Christensen, Reference Christensen1957; Hill and Lacey, Reference Hill and Lacey1983).
Germination cycles
Dormancy cycles occur in seeds of many species with non-deep PD, and they can be explained by seasonal changes in the physiology and molecular biology of the seeds (Courtney, Reference Courtney1968; Taylorson, Reference Taylorson1970; Baskin and Baskin, Reference Baskin and Baskin1985, Reference Baskin and Baskin2014a; Footitt et al., Reference Footitt, Douterelo-Soler, Clay and Finch-Savage2011). During dormancy cycling, seeds cycle between dormancy ↔ non-dormancy or between conditional dormancy ↔ non-dormancy, and during the non-dormant period they can germinate if exposed to favourable environmental conditions. However, if non-dormant seeds are faced with unfavourable conditions they are induced into secondary dormancy (Baskin and Baskin, Reference Baskin and Baskin2014a). We totally agree with Paulsen et al. (Reference Paulsen, Colville, Kranner, Daws, Hogstedt, Vandvik and Thompson2013) that, in contrast to seeds with PD, dormancy in those with PY can only be broken once, i.e. seeds with PY do not cycle between dormancy and non-dormancy. However, we do not agree that lack of a dormancy cycle ‘… raises questions about its [PY] efficiency as a dormancy trait’. They do not mention sensitivity cycling in seeds with PY, which is discussed below.
Cyclic germination patterns have been observed in seeds with PY mainly in Fabaceae and Convolvulaceae (Taylor and Revell, Reference Taylor and Revell1999; Van Assche et al., Reference Van Assche, Debucquoy and Rommens2003; Benvenuti et al., Reference Benvenuti, Dinelli, Bonetti and Catizone2005; Jayasuriya et al., Reference Jayasuriya, Baskin and Baskin2008a, Reference Jayasuriya, Baskin, Geneve and Baskinb) and are related to sensitivity cycling (Jayasuriya et al. Reference Jayasuriya, Baskin and Baskin2008a, Reference Jayasuriya, Baskin and Baskin2009a). Seeds with PY cannot cycle between dormancy ↔ non-dormancy since dormancy break occurs via a physical break in the seed coat. This physical break cannot be resealed (Jayasuriya et al., Reference Jayasuriya, Baskin, Geneve and Baskin2007). Sensitive seeds become non-dormant if they are exposed to the environmental conditions that can break dormancy; otherwise, if sensitive seeds are exposed to unfavourable conditions they re-enter the insensitive state. Seeds in the insensitive state are not sensitive to dormancy breaking cues, and insensitive seeds have to be exposed to certain environmental conditions to become sensitive again (Jayasuriya et al., Reference Jayasuriya, Baskin and Baskin2009a). Thus, sensitivity cycling of seeds with PY serves the same ecological function as dormancy cycling in seeds with PD.
Seed longevity in soil seed banks
The longevity of a buried seed varies with the original vigour of the seed, depth of burial and environmental conditions during burial (Darlington and Steinbauer, Reference Darlington and Steinbauer1961). Seed-coat characters such as thickness and permeability and seed surface:volume ratio directly influence the longevity, germination percentage and rate (speed) and persistence of seeds in the seed bank (Thompson et al., Reference Thompson, Grime and Mason1977, Reference Thompson, Band and Hodgson1993; Leck, Reference Leck, Leck, Parker and Simpson1989; Chambers, Reference Chambers1995). Further, Darlington and Steinbauer (Reference Darlington and Steinbauer1961) mentioned that the longevity of buried seed is related to amount of water absorbed by the seed from the environment. Thus, the water-impermeable seed coat prevents movement of water into the seed and thereby assists in retention of viability.
The classic example of a species with long-lived seeds is the Indian lotus, Nelumbo nucifera (1288 ± 271 years old) (Shen-Miller et al., Reference Shen-Miller, Mudgett, Schopf, Clarke and Berger1995), which has a water-impermeable fruit coat. These authors suggested that the Indian lotus can survive for more than 1000 years due to the water- and air-tight coat. Physically dormant seeds of Malva rotundifolia were viable after 120 years in the Beal buried-seed experiment (Telewski and Zeevaart, Reference Telewski and Zeevaart2002) and I. lacunosa for 39 years (when the experiment was terminated) in the Duvel buried-seed experiment (Toole and Brown, Reference Toole and Brown1946). However, PY is not a necessity for seeds to have a long life span. Seeds of Verbascum sp. and V. blattaria (both with PD) were viable after 120 years, and those of Oenothera biennis and of Rumex crispus (both with PD) were viable after 80 years of burial in Dr Beal's experiment (Telewski and Zeevaart, Reference Telewski and Zeevaart2002). Seeds of wild date palm (Phoenix dactylifera), which have morphological or morphophysiological dormancy (Baskin and Baskin, Reference Baskin and Baskin2014b) may hold the record for length of period for retention of viability. Viable germinable seeds of date palm found in excavations of Masada (an Herodian fortress) overlooking the Dead Sea (Israel) were radiocarbon dated at 297 bc, making them about 2000 calendar years old (Sallon et al., Reference Sallon, Solowey, Cohen, Korchinsky, Egli, Woodhatch, Simchoni and Kislev2008). Wild date palm seeds from the same sample were germinated successfully and gave rise to healthy seedlings.
Concluding remarks
We have presented strong evidence that not only predator pressure but also several other environmental pressures were likely involved in the evolution of PY. Therefore, the crypsis hypothesis of Paulsen et al. (Reference Paulsen, Colville, Kranner, Daws, Hogstedt, Vandvik and Thompson2013) reveals only one of the adaptive roles of PY. Further, their argument for evolution of the water-impermeable seed coat in response to predator pressure is not consistent with Darwin's (Reference Darwin1859) theory of evolution by natural selection. Nor is it consistent with the multiple times that PY has evolved since the origin of the angiosperms, or with the occurrence of PY in all the major terrestrial biomes and in wet, dry, saline and non-saline habitats. Further, PD is more common than PY in hot deserts, where one would expect to find the highest selection pressure from seed-eating rodents (Baskin and Baskin, Reference Baskin and Baskin2014a).
Conflicts of interest
None.
