Introduction
‘Seed (diaspore) heteromorphism’ is a broadly- and loosely-defined term used to describe differences in size/mass, morphology, position on mother plant and ecological function (e.g. dispersal, dormancy/germination) of two or more seeds or other diaspores produced by an individual plant. Thus, the term is applied to a variety of situations concerning degree of distinctness of differences in size/mass, morphology and position of diaspores on a plant. For example, Aethionema arabicum (Brassicaceae) produces two morphologically distinct aerial diaspores with no intermediates (Arshad et al., Reference Arshad, Sperber, Steinbrecher, Nichols, Jansen, Leubner-Metzger and Mummenhoff2019), Heterosperma pinnatum (Asteraceae) produces two morphologically distinct diaspores connected by a series of morphologically intermediate ones (Venable et al., Reference Venable, Búrquez, Corral, Morales and Espinosa1987, Reference Venable, Dyreson and Morales1995; Martorell and Martínez-López, Reference Martorell and Martínez-López2014) and Ceratocarpus arenarius (Amaranthaceae) produces two morphologically distinct ground-level diaspores and a series of aerial diaspores that differ continuously in size and morphology from top to bottom of the plant canopy (Lu et al., Reference Lu, Tan, Baskin and Baskin2013). It is no wonder, then, that in a recent paper Scholl et al. (Reference Scholl, Calle, Miller and Venable2020) stated that defining seed heteromorphism is a challenge.
The aim of this review paper is to provide an in-depth classification scheme, an earlier version of which was published in the second edition of ‘Seeds’ by Baskin and Baskin (Reference Baskin and Baskin2014), based on size, morphology and position on the mother plant that will give more exactness to use of the term ‘seed heteromorphism’. More generally, our aim was to characterize the diversity of structural design (‘architecture’) of diaspore monomorphism and heteromorphism in angiosperms.
Methods
Our classification scheme is based on information in the literature on the size/mass, morphology and position (e.g. aerial, basal and subterranean) on the mother plants of the diaspores of angiosperm taxa. Basically, we greatly expanded/modified the diaspore classification schemes of Mandák (Reference Mandák1997) and Barker (Reference Barker2005), sometimes using terminology that Zohary (Reference Zohary1937, Reference Zohary1962), Ellner and Shmida (Reference Ellner and Shmida1981) and van der Pijl (Reference van der Pijl1982) applied to diaspore dispersal. Seeds (diaspores) were first divided into two major categories (monomorphic and heteromorphic) called divisions and each Division into several successively lower hierarchical layers.
Our scheme does not include the terms that Zohary (Reference Zohary1937, Reference Zohary1962), Ellner and Shmida (Reference Ellner and Shmida1981), van der Pijl (Reference van der Pijl1982) and/or Gutterman (Reference Gutterman1993, Reference Gutterman1994b) used to describe agents/modes of dispersal such as anemochory (wind), hydrochory (water), ombrohydrochory (rain) and zoochory (animals). Neither does it include genetic polymorphism, in which (1) two or more kinds of diaspores are produced by a species, (2) a population may be diaspore-monomorphic or dimorphic/polymorphic and (3) individual plants produce only one kind of diaspore, which is determined by Mendelian inheritance, i.e. a genetic segregation, not a somatic differentiation (see Appendix A).
In general, we follow Scholl et al.'s (Reference Scholl, Calle, Miller and Venable2020) definition of ‘seed heteromorphism.’ According to these authors, variation in the morphology of heteromorphic diaspores can be discrete or continuous. However, if the variation is continuous the most extreme diaspores need to be widely divergent morphologically, such as occurs in H. pinnatum (see Introduction), for them to qualify as heteromorphic. Diaspores that do not meet one of these two criteria are, by default, considered to be monomorphic.
Results and discussion
The classification scheme we assembled for diaspore monomorphism and heteromorphism is shown in Table 1. However, as pointed out by Imbert (Reference Imbert2002) and Scholl et al. (Reference Scholl, Calle, Miller and Venable2020) seed heteromorphism (and thus monomorphism) is not easy to define. According to Harper et al. (Reference Harper, Lovell and Moore1970), most individual plants and populations have a normal or skewed (continuous) distribution for seed size or shape, but that ‘It is, however, characteristic of some species to produce two or more sharply defined types of seed.’ Thus, the word ‘monomorphic’ does not mean that all seeds on an individual plant have a single size/mass or morphology (or that they have the same degree of dormancy). In fact, there is considerable variation in these traits (especially mass) among monomorphic seeds on an individual plant and even among those on the same infructescence or within a fruit as various authors have reported (e.g. Janzen, Reference Janzen1977a, Reference Janzenb; Ernst, Reference Ernst1981; Thompson, Reference Thompson1981; Pitelka et al., Reference Pitelka, Thayer and Hansen1983; Stanton, Reference Stanton1984; Thompson, Reference Thompson1984; Wolf et al., Reference Wolf, Hainsworth, Mercier and Benjamin1986; Wulff, Reference Wulff1986; Michaels et al., Reference Michaels, Benner, Hartgerink, Lee and Rice1988; McGinley, Reference McGinley1989; Winn, Reference Winn1991; Fenner, Reference Fenner1992; Susko and Lovett-Doust, Reference Susko and Lovett-Doust2000). Monomorphic simply means that seeds cannot easily be sorted into two or more clearly-defined (distinct) groups based on traits such as size/mass and/or morphology. Monomorphic seeds may show differences in dormancy/germination: cryptic polymorphism (i.e. ecological differentiation in the absence of obvious morphological differences) of Venable (Reference Venable1985). Heteromorphic seeds, on the other hand, means that seeds on an individual plant readily can be sorted into two or more distinct groups that differ in size/mass and/or morphology.
An excellent example of cryptic seed heteromorphism has been reported by Liyanage et al. (Reference Liyanage, Ayre and Ooi2016) for two species of Fabaceae (Bossiaea heterophylla and Viminaria juncea) whose seeds have physical dormancy. They found that individual plants of the two species, which occur in fire-prone ecosystems in southeastern Australia, produce seeds with a high threshold temperature and a low threshold temperature for physical dormancy break. There were no significant differences in mass or visible differences in shape or color of high- and low threshold temperature seeds from individual plants of either species. Seeds with a low threshold temperature for dormancy break could germinate after exposure to temperatures of a low-intensity fire (40–60°C), whereas those with a high threshold temperature for dormancy break could germinate only after exposure to temperatures of a high-intensity fire (≥80°C). Further, under competition with seedlings of Acacia linifolia (Fabaceae), a co-occurring species, seedlings of B. heterophylla emerging after low-intensity fire temperature grew better than those emerging after high-intensity fire temperature. Competition would be more intense following low- than high-intensity fire due to the survival of more plants in the low- than high-intensity-burned community. Thus, although low and high threshold temperature seeds did not exhibit differences in mass or morphology they differed in their dormancy-breaking response to fire intensity and in seedling growth response to competition resulting from different fire intensities.
Overall, we show that there is considerable structural diversity in design (‘architecture’) of diaspore-heteromorphic angiosperm species based on diaspore size, morphology and position on the mother plant and of diaspore-monomorphic angiosperms based on diaspore position (e.g. aerial, basal and subterranean) on the mother plant. To date, only a few hundred of the >250,000 extant angiosperm species have been reported to produce heteromorphic diaspores (Imbert, Reference Imbert2002; Wang et al., Reference Wang, Tan, Baskin and Baskin2010; Baskin et al., Reference Baskin, Lu, Baskin, Tan and Wang2014; Scholl et al., Reference Scholl, Calle, Miller and Venable2020; Zhang et al., Reference Zhang, Baskin, Baskin, Cheplick, Yang and Huang2020), apparently meaning that the vast majority of flowering plants produce monomorphic diaspores.
In a recent survey for homomorphic (=our monomorphic) and heteromorphic species in the North American deserts, using information in various floras for the area, Scholl et al. (Reference Scholl, Calle, Miller and Venable2020) identified 458 monomorphic species and 101 heteromorphic taxa, of which 75 of the latter were annuals. They also reported that their study brought the total number of known seed heteromorphic species to 378. The flora of this area contains many annuals, the climate is arid/semiarid and amount and timing of rainfall is unpredictable, which are conditions that favor bet-hedging via diaspore heteromorphism as a life history strategy (Scholl et al., Reference Scholl, Calle, Miller and Venable2020; Gianella et al., (Reference Gianella, Bradford and Guzzon2021). Thus, the North American deserts undoubtedly are highly over-represented in proportion of heteromorphic species compared to other bioclimatic regions. In which case, we should expect that the proportion of diaspore-heteromorphic species in the world's flora is much lower than that in the North American deserts.
Mandák's (Reference Mandák1997) classification scheme for seed (diaspore) heteromorphism divides heteromorphic diaspores into Amphicarpy and Heterodiaspory and distinguishes three subgroups for the latter category: heterocarpy (Heteromericarpy of van der Pijl, Reference van der Pijl1982), heteroarthrocarpy and heterospermy. (See Table 1 for definition of each of these three terms and of those mentioned in the following.) Barker's (Reference Barker2005) diaspore classification scheme deals only with basicarpy, geocarpy and amphicarpy. His scheme includes full geocarpy, with three subtypes, i.e. active geocarpy, geophytic geocarpy and passive geocarpy; and basicarpy, also with three subtypes, i.e. aerial amphicarpy, amphi-geocarpy and amphi-basicarpy. Note that all of the categories in Mandák's (Reference Mandák1997) scheme fit under our Division II (Heteromorphic), whereas the scheme of Barker (Reference Barker2005) includes terms under both our Division I (e.g. Full geocarpy) and Division II (e.g. Amphi-basicarpy). Imbert (Reference Imbert2002) recognized two categories of diaspore heteromorphism: heterocarpy and heterospermy.
Many species of grasses (Poaceae) produce cleistogamous (CL) axillary spikelets within leaf sheaths at nodes on flowering culms (Connor, Reference Connor1979, Reference Connor1981; Campbell et al., Reference Campbell, Quinn, Cheplick and Bell1983). Some grasses, e.g. the well-studied species Triplaris purpurea (e.g. Chase, Reference Chase1908, Reference Chase1918; Cheplick, Reference Cheplick1996; Cheplick and Sung, Reference Cheplick and Sung1998), produce these clandestine spikelets at all nodes on the flowering culm, which is terminated by an inflorescence of chasmogamous (CH) spikelets. Caryopsis mass decreases continuously [in a log-linear (or nearly so) fashion] from the lowermost to one of the upper leaf sheaths, above which there is little or no change in mass of caryopses, including those in the terminal CH spikelets. Chase (Reference Chase1908, Reference Chase1918) used the term ‘cleistogene’ to describe the solitary sessile single floret with palea and lemma but without glumes in the lower leaf sheaths of T. purpurea. Chase (Reference Chase1918) applied the term ‘chasmogene’ to the terminal (‘ordinary’) spikelet. Her illustrations clearly show that the caryopsis from the cleistogene is much larger than that from the chasmogene. We have (cautiously) suggested that T. pupurea might fit subgroup B of amphi-basicarpy (Table 1).
Monomorphic aerial CH/CL plants/populations may produce only CH or only CL flowers (Wilken, Reference Wilken1982; Sun, Reference Sun1999; Hayamizu et al., Reference Hayamizu, Hosokawa, Kimura and Ohara2014). Corallorhiza bentleyi (Orchidaceae), is an example of a species in which some populations produce only CL and others both CH and CL (see Freudenstein, Reference Freudenstein1999; Culley and Klooster, Reference Culley and Klooster2007). Small or young individuals of some CL species or individuals growing under unfavorable conditions produce only CL (Coker, Reference Coker1962; Minter and Lord, Reference Minter and Lord1983; Oakley et al., Reference Oakley, Moriuchi and Winn2007; Hayamizu et al., Reference Hayamizu, Hosokawa, Kimura and Ohara2014). Epifagus virginiana (Orobanchaceae) is an annual CL holoparasite with ‘dust seeds’ and an undifferentiated embryo that is host-specific on the roots of American beech (Fagus grandifolia). However, most, and sometimes all, of the flowers are CL, and the CH flowers usually are sterile. Thus, most seeds are produced by CL flowers (Schrenk, Reference Schrenk1894; Cooke and Schively, Reference Cooke and Schively1904; Thieret, Reference Thieret1969, Reference Thieret1971; Musselman, Reference Musselman1982).
Lloyd and Schoen (Reference Lloyd and Schoen1992) state that CH and CL seeds in most CL species differ in size, dispersal and germination. However, seeds from CH and CL flowers of our Group B of Supergroup 1 (monomorphic aerial) may or may not differ in these respects (Baskin and Baskin, Reference Baskin and Baskin2014, Reference Baskin and Baskin2017). When there is a difference in mass of CH and CL seeds, the mass of CH seeds usually is greater than that of CL seeds (e.g. Cope, Reference Cope1966; Hiratsuka and Inoue, Reference Hiratsuka and Inoue1988; Cheplick, Reference Cheplick2005; Eckstein et al., Reference Eckstein, Hölzel and Danihelka2006; Albert et al., Reference Albert, Campbell and Whitney2011; Huebner, Reference Huebner2011; Munguía-Rosas et al., Reference Munguía-Rosas, Parra-Tabla, Ollerton and Cervera2012). Differences in germination of CH and CL diaspores are more likely to occur in amphicarpic sensu stricto species than in aerial CL species. In 58 of 65 case studies (89.2%) for amphicarpic sensu stricto species, seeds from CH and Cl flowers differed in germination percentage, whereas in aerial CL species 83 of 132 case studies (62.9%) of seeds from CH and Cl flowers differed in germination percentage (Baskin and Baskin, Reference Baskin and Baskin2017). Further study may show that a new category needs to be split out of Group B (Supergroup I, Division I) and incorporated into Division II. The new category would include species in which fruits/seeds produced by CH and CL flowers are morphologically distinct and (presumably) differ in ecology.
In Ceratocarpus arenarius, the basicarps differ in morphology and mass (and in embryo mass) from the aerial dispersal/germination units, which show continuous variation (increase or decrease) in morphology (see Supplementary Table S1 in Lu et al., Reference Lu, Tan, Baskin and Baskin2013). Thus, there is discontinuous variation in this species between the basicarps (a) and aerial dispersal units (b)–(f) (Lu et al., Reference Lu, Tan, Baskin and Baskin2013). Gao and Wei (Reference Gao and Wei2007) and Gao et al., (Reference Gao, Wei and Yan2008) recognized only two morphological types of fruits on plants of this species, namely subterranean (the two basicarps, which are, in fact, basal and not subterranean) and aerial.
In Pisum fulvum (Fabaceae), there is a gradient from amphicarpic plants (sensu stricto) with both aerial and subterranean flowers and fruits to plants that produce only aerial flowers and fruits (Mattatia, Reference Mattatia1977). One of the stages in the gradient is a basicarpic form that produces CH flowers near the soil surface, which Mattatia (Reference Mattatia1977) called ‘subamphicarpic.’ Thus, this species consists of both monomorphic (e.g. basicarpic form) and heteromorphic (e.g. amphicarpic form) plants.
Some species may exhibit plasticity as to the diaspore classification category to which they belong. For example, species that are amphicarpic and produce both aerial and subterranean diaspores under favorable environmental conditions may produce only underground fruits under stressful conditions and thus be ‘facultatively geocarpic’, e.g. Amphicarpaea edgeworthii (Fabaceae) (Zhang et al., Reference Zhang, Baskin, Baskin, Yang and Huang2017), Amphicarpum amphicarpon (Poaceae) Cheplick and Quinn, Reference Cheplick and Quinn1983) and Gymnarrhena micrantha (Asteraceae) (Koller and Roth, Reference Koller and Roth1964; Zeide, Reference Zeide1978; Loria and Noy-Meir, Reference Loria and Noy-Meir1979/80). In which cases, individuals of these two annual species have the capacity to be either diaspore-heteromorphic (amphicarpic) or monomorphic (geocarpic).
Unlike these amphicarpic species, the cold desert annual diaspore-polymorphic amphi-basicarpic species Ceratocarpus arenarius produces both basal (typically two) and aerial diaspores in the most stressful conditions in which it grows in its cold desert habitat. In the cold deserts of northwest China, we have observed that the smallest (5 cm tall, no branches) and largest (35 cm tall, bushy) plants of this species produce both basal and aerial diaspores, albeit in different basal morph:aerial morph and within-aerial morph ratios. Detailed experimental garden studies on the effect of abiotic (e.g. soil physicochemistry) and biotic (i.e. inter- and intraspecific competition) stress on phenotypic plasticity of the growth and reproduction of C. arenarius, including variation in diaspore morph ratios, have been published by Gan et al. (Reference Gan, Lu, Baskin, Baskin and Tan2020) and Lu et al. (Reference Lu, Gan, Tan, Baskin and Baskin2021).
Concluding remarks
We have documented via a hierarchical-based classification system the considerable diversity in structural design (‘architecture’) of diaspore monomorphism and heteromorphism in angiosperms. The scheme will enable investigators working on the broad topic of ‘seed heteromorphism’ to more precisely communicate their research to others, in part at least by giving more exactness to the term. It also will aid plant biologists in the preparation of a profile (spectrum) of the kinds (hierarchical categories) of diaspore monomorphism and heteromorphism for the various ecological and biogeographical units on earth. Finally, a detailed classification scheme that includes both diaspore monomorphism and hetermorphism is required for addressing the phylogenetic/evolutionary aspects of ‘seed heteromorphism’ in angiosperms (e.g. see Fernández et al., Reference Fernández, Aguilar, Panero and Feliner2001; Cruz-Mazo et al., Reference Cruz-Mazo, Buide, Samuel and Narbona2009, Reference Cruz-Mazo, Narbona and Buide2010). All that being said, however, it is likely that more hierarchical categories will need to be added to our system and/or existing ones revised/refined as literature and field research continues on diaspore monomorphism and heteromorphism in angiosperms.
Conflict of interest
The authors declare that they have no competing interests.
Appendix A Genetic diaspore polymorphism in angiosperms
Genetic diaspore polymorphism is best known in the family Valerianaceae, including Plectritis (Ganders et al., Reference Ganders, Carey and Griffiths1977a,Reference Ganders, Carey and Griffithsb; Carey and Ganders, Reference Carey and Ganders1980) and Valerianella (Eggers Ware, Reference Eggers Ware1983). In Plectritis congesta, experimental crosses showed that the kind of fruit morph is monogenically inherited with the allele for winged fruits dominant over wingless fruits (Ganders et al., Reference Ganders, Carey and Griffiths1977a). In Plectritis brachystemon, homozygous winged plants (i.e. selfed plants produced only winged fruits) x homozygous wingless fruits (i.e. selfed plants produced only wingless fruits) → F1 hybrids, all of which produced winged fruits. The F2 (F1 selfed) consisted of 27 plants that produced winged fruits and nine that produced wingless fruits (a 3:1 ratio). Thus, fruit dimorphism in P. brachystemon is controlled by a single locus with the allele for winged fruit dominant (Ganders et al., Reference Ganders, Carey and Griffiths1977b). In Valerianella ozarkana, crosses between forma ozarkana (winged) and forma bushii (fusiform) indicated that there was a simple monogenic relationship in which the ozarkana (winged) allele is dominant over the bushii (fusiform) allele (Eggers-Ware, Reference Eggers Ware1983). It should be pointed out that not all species that produce two or more kinds of diaspores fit either somatic polymorphism or genetic polymorphism, i.e. they neither fit one nor the other. Fedia cornucopiae and F. grandflora (Valerianaceae) have been shown to exhibit both somatic and genetic polymorphism (Mathez and Xena de Enrich, Reference Mathez, Xena de Enrich, Jacquar, Heim and Antonovics1985a, Reference Mathez and Xena de Enrich1985b).,
Genetic diaspore heteromorphism (only one morph per plant) has been reported in the two grass species Aegilops speltoides (e.g. Zohary and Imber, Reference Zohary and Imber1963; Belyayev and Raskina, Reference Belyayev and Raskina2013; Ruban and Badaeva, Reference Ruban and Badaeva2018) and Achnatherum hymenoides (Jones and Nielson, Reference Jones and Nielson1999; Jones et al., Reference Jones, Redinbaugh, Larson, Zhang and Dow2007). In one of the two morphotypes of A. speltoides (i.e. form aucheri), the spike is the dispersal unit, and in the other morphotype (form ligustica) the spikelet is the dispersal unit. In A. hymenoides, seed size, i.e. jumbo > globose > elongate) is under genetic control, and degree of seed dormancy decreases from jumbo to globose to elongate.
A kind of genetic polymorphism for capitulum type occurs in British populations of Senecio vulgaris (Asteraceae). Plants of this species have radiate and non-radiate capitulum morphs, which are under genetic control. The radiate morph originated via introgressive hybridization between the native non-radiate allotetraploid S. vulgaris and the introduced diploid radiate S. squalidus. Two tightly-linked genes of S. squalidus were introgressed into S. vulgaris. A short-rayed form (a heterozygote) is produced from crosses between radiate and non-radiate forms of S. vulgaris. Fresh cypselae from radiate and non-radiate morphs differ in germination phenology. There also were differences in germination percentages after cypselae from radiate and non-radiate morphs were stored in the laboratory from October to June, during which time after-ripening may have occurred (e.g. Trow, Reference Trow1912; Richards, Reference Richards1975; Ingram et al., Reference Ingram, Weir and Abbott1980; Abbott, Reference Abbott1986, Reference Abbott, Ashton and Forbes1992; Abbott et al., Reference Abbott, Horrill and Noble1988, Reference Abbott, Ashton and Forbes1992, Reference Abbott, Bretagnolle and Thébaud1998; Abbott and Horrill, Reference Abbott and Horrill1991; Chapman and Abbott, Reference Chapman and Abbott2010).
Spergula and Spergularia are two genera in the Caryophyllaceae containing species that produce heteromorphic seeds with a genetic component. Spergula arvensis produces papilate (P) and smooth, i.e. non-papilate (NP), seeds and a hybrid (P × NP) intermediate between the two morphs; the hybrid is produced in a low frequency. Inheritance of seed coat character is controlled by one gene, one locus and two alleles. The intermediate morph is heterozygous with incomplete dominance (New, Reference New1958, Reference New1959, Reference New1961, Reference New1978; New and Herriott, Reference New and Herriott1981; Wagner, Reference Wagner1986, Reference Wagner1988; Kucewicz and Gojło, Reference Kucewicz and Gojło2013).
Several species of Spergularia (e.g. S. marina and S. media) produce winged and non-winged seeds and a heterzygous intermediate with a narrow wing produced in a low frequency (Salisbury, Reference Salisbury1958; Sterk, Reference Sterk1969a,Reference Sterkb,Reference Sterkc,Reference Sterkd; Sterk and Dijkhuizen, Reference Sterk and Dijkhuizen1972; Telenius and Torstensson, Reference Telenius and Torstensson1989, Reference Telenius and Torstensson1991; Telenius, Reference Telenius1992; Ceynowa-Giełdon, Reference Ceynowa-Giełdon1993; Redbo-Torstensson, Reference Redbo-Torstensson1994; Redbo-Torstensso and Telenius, Reference Redbo-Torstensson and Telenius1995; Reference Telenius and Torstensson1999; Mazer and Lowrey, Reference Mazer and Lowrey2003; Memon et al., Reference Memon, Bhatti, Khalid, Arshad, Mirbahar and Qureshi2010). Seeds of Spergularia diandra collected near Sede Boker in the Negev Desert of Israel consisted of three genotypes (hairy, partly-hairy and smooth). Each genotype had three seed-color phenotypes (black, brown and yellow), and there were differences in germination of the phenotypes. Thus, there were nine seed morphs, i.e. 3 phenotypes × 3 genotypes = 9 types of seed morphs (Gutterman, Reference Gutterman1994a, Reference Gutterman1996, Reference Gutterman1997a, Reference Gutterman, Ellis, Black, Murdoch and Hongb, Reference Gutterman and Ambasht1998, Reference Gutterman2000).