Results and discussion

Palm fruits are berries or fibrous drupes with up to ten seeds. However, in many species there is only one seed per mature fruit, although the ovary is tricarpellate. The fruits exhibit a wide range of sizes, the largest being that of the double coconut Lodoicea maldivica (c. 45–50 cm long and weighing 36 kg), which has the largest seed in the plant kingdom. Generally, the fruit wall (pericarp) consists of the exocarp, mesocarp and endocarp. The endocarp or ‘shell’ may be thick and stoney (drupe) to thin, papery or fleshy (berry); in some taxa, it is undifferentiated (Corner, Reference Corner1966; Harper et al., Reference Harper, Lovell and Moore1970; Moore and Uhl, Reference Moore and Uhl1973; Uhl and Dransfield, Reference Uhl and Dransfield1987; Tomlinson, Reference Tomlinson1990; Dransfield and Uhl, Reference Dransfield, Uhl and Kubitzki1998).

The seed consists of an achlorophyllous embryo and endosperm surrounded by a thin (usually) seed coat and sometimes a sarcotesta (fleshy outer seed coat). In most species, the embryo is small and linear, but in some species, e.g. Nypa fruticosa (see figure B.51 in Tomlinson, Reference Tomlinson1986), the only palm that is a ‘strict or true mangrove species’, it is shaped like a cylinder, being flat on one end and tapering to a point on the other end. In general, the embryo (E):seed (S) length (E:S) ratio is low, i.e. < 0.10 to c. 0.30–0.40. However, in a few genera the embryo is large, e.g. E:S ratio of c. 0.70 in Oenocarpus and c. 0.85 in Jessenia [E:S ratios determined by measuring lengths of seeds and embryos from drawings in Uhl and Dransfield (Reference Uhl and Dransfield1987)]. The embryo may occupy < 1% of the total seed volume, and its position can be basal (micropylar end), apical (other end), lateral or some variation thereof, e.g. subapical. In seeds with thick endocarps, such as those of coconut palm and oil palm, the embryo is located adjacent to an endocarp germination pore (‘eye’) (Fig. 1a). Seed storage behaviour can be orthodox or non-orthodox (intermediate or recalcitrant) (Corner, Reference Corner1966; Yakovlev and Zhukova, Reference Yakovlev and Zhukova1980; Janzen, Reference Janzen1982; Uhl and Dransfield, Reference Uhl and Dransfield1987; Tomlinson, Reference Tomlinson1990; Ellis et al., Reference Ellis, Hong, Roberts and Soetisna1991; Orozco-Segovia et al., Reference Orozco-Segovia, Batis, Rojas-Arechiga and Mendoza2003; Panza et al., Reference Panza, Lainez and Maldonado2004; Pritchard et al., Reference Pritchard, Wood, Hodges and Vautier2004; Von Fintel et al., Reference Von Fintel, Berjak and Pammenter2004; González-Benito et al., Reference González-Benito, Huertas-Micó and Pérez-Garcia2006; Perez et al., Reference Perez, Criley and Baskin2008; Ribeiro et al., Reference Ribeiro, Souza, Rodrigues, Oliveira and Garcia2011, Reference Ribeiro, Oliveira and Garcia2012a, b; Jose et al., Reference Jose, Erasmo and Coutinho2012).

Figure 1 (a) Longitudinal section of a palm fruit showing embryo with cotyledonary petiole adjacent to the operculum (micropylar endosperm + seed coat) under the germination pore of a drupe of a palm. The embryo is about 3 mm long (from Hussey, Reference Hussey1958). (b) Median longitudinal section of proximal region of cotyledonary petiole showing embryonic axis (ep, ra, rc) that is enclosed within it. Scale bar is 0.20 mm (from Panza et al., Reference Panza, Lainez and Maldonado2004). (c) The three primary morphological types of seedlings (germination) commonly recognized in palms (from Tomlinson, Reference Tomlinson1960). Aa, Ab, Ac, remote-tubular; B, remote-ligular; Ca, Cb, Cc, adjacent-ligular. CP, cotyledonary petiole; EC, endocarp; ep, epicotyl; ES, endosperm; GP, germination pore with fibrous mesocarp tissue; HA, haustorium; ME, micropylar endosperm (endosperm cap); ra, radicle; rc, root cap; SC, seed coat. Figures reproduced with permission.

In whole-view (Fig. 1a), the embryo can be divided into two regions: cotyledonary petiole located adjacent to the operculum [portion of endosperm (micropylar endosperm) and seed coat covering the proximal end of embryo that may be demarcated by a zone of weakness] and cotyledon blade, which remains inside the seed. The endosperm is homogeneous or ruminate, and in some species it contains a cavity that is dry or filled with liquid, e.g. coconut ‘milk’ in Cocus nucifera (Hussey, Reference Hussey1958; Corner, Reference Corner1966; Robertson, Reference Robertson1977; DeMason, Reference DeMason1985; Uhl and Dransfield, Reference Uhl and Dransfield1987; DeMason et al., Reference DeMason, Stillman and Ellmore1989; Aguiar and de Mendonça, Reference Aguiar and de Mendonça2002; Neves et al., Reference Neves, Ribeiro, Cunha, Pimenta, Mercadante-Simoes and Lopes2013). The cotyledonary blade or apical portion thereof acts as a haustorium and absorbs and metabolizes products of hydrolysis from the endosperm; however, see Panza et al. (Reference Panza, Lainez and Maldonado2004), who concluded that the endosperm of the recalcitrant species Euterpe edulis is inactive, i.e. no storage metabolism. Histologically, the embryo consists of a single cotyledon and a short plumular–radicular axis (epicotyl with leaf primordia and root apex) located in the proximal end of the cotyledonary tube (Fig. 1b) (Dassanayake and Sivakadachchan, Reference Dassanayake and Sivakadachchan1973; Robertson, Reference Robertson1976; Haccius and Philip, Reference Haccius and Philip1979; DeMason, Reference DeMason1988a; Aguiar and de Mendonça, Reference Aguiar and de Mendonça2003; Panza et al., Reference Panza, Lainez and Maldonado2004; Henderson, Reference Henderson2006; Ribeiro et al., Reference Ribeiro, Oliveira and Garcia2012a; Nazario et al., Reference Nazario, do Nascimento Ferreira, de Lima e Borges, Genovese-Marcomíni and de Mendonça2013; Neves et al., Reference Neves, Ribeiro, Cunha, Pimenta, Mercadante-Simoes and Lopes2013).

In some palms, the natural dispersal unit is a drupe with a hard, thick endocarp from which the seeds can be removed only by physically cracking it (Robertson and Small, Reference Robertson and Small1977; Broschat, Reference Broschat1998; Davies and Pritchard, Reference Davies and Pritchard1998; Moussa et al., Reference Moussa, Margolis, Dube and Odongo1998; Ehara et al., Reference Ehara, Morita, Komada and Goto2001; Perez et al., Reference Perez, Criley and Baskin2008; Myint et al., Reference Myint, Chanprasert and Srikul2010; Ribeiro et al., Reference Ribeiro, Souza, Rodrigues, Oliveira and Garcia2011; Neves et al., Reference Neves, Ribeiro, Cunha, Pimenta, Mercadante-Simoes and Lopes2013). In most cases, presence of a hard endocarp inhibits seed germination. Thus, isolated seeds (removed from endocarp) germinate better, often much better (higher percentage/rate), than those with the pericarp intact (Broschat, Reference Broschat1998; Moussa et al., Reference Moussa, Margolis, Dube and Odongo1998; Ferreira and Gentil, Reference Ferreira and Gentil2006; Perez et al., Reference Perez, Criley and Baskin2008; Ribeiro et al., Reference Ribeiro, Souza, Rodrigues, Oliveira and Garcia2011). As such, claims have been made that the fruit is water impermeable, thus accounting for the ‘physical dormancy’ of the diaspore, i.e. drupe (Carpenter et al., Reference Carpenter, Ostmark and Cornell1993; Moussa et al., Reference Moussa, Margolis, Dube and Odongo1998). However, this is not the case. Robertson and Small (Reference Robertson and Small1977) showed that the hard, thick pericarp of Jubaeopsis caffra was permeable to water (and oxygen), and Perez et al. (Reference Perez, Criley and Baskin2008) demonstrated that imbibition of seeds inside fruits of Pritchardia remota was not blocked. The water content of isolated seeds of Acrocomia aculeata did not differ significantly from that of intact fruits with a thick, stony endocarp (21.7 and 22.9%, respectively) after 60 d on a moist substrate (Ribeiro et al., Reference Ribeiro, Souza, Rodrigues, Oliveira and Garcia2011). After 45 d of imbibition, the percentage moisture content (MC) of isolated seeds, seeds in intact fruits and seeds inside scarified fruits of Attalea vitrivir was nearly identical; seed MC increased from c. 6–7% to c. 22–23% (Neves et al., Reference Neves, Ribeiro, Cunha, Pimenta, Mercadante-Simoes and Lopes2013).

We consider the embryo in palm seeds to be underdeveloped. The proximal end of the embryo, i.e. the cotyledonary petiole, is adjacent to the operculum, and germination of palm seeds is often defined as dislodgment of the operculum resulting from elongation of the cotyledonary petiole (Fong, Reference Fong1978; Perez et al., Reference Perez, Criley and Baskin2008; Fior et al., Reference Fior, Rodrigues, Leonhardt and Schwarz2011; Ribeiro et al., Reference Ribeiro, Souza, Rodrigues, Oliveira and Garcia2011; Neves et al., Reference Neves, Ribeiro, Cunha, Pimenta, Mercadante-Simoes and Lopes2013). Further, the cotyledonary blade (haustorium) expands to fill most of the seed, albeit after the cotyledonary petiole emerges from the seed (germination), i.e. expansion of the cotyledon/haustorium occurs and tissue development continues following germination (DeMason, Reference DeMason1984, Reference DeMason1985, Reference DeMason1988b). In which case, germination occurs before emergence of the radicle and shoot from the cotyledonary petiole/sheath and after the latter has protruded through the opening left by displacement of the operculum. Thus, the embryo requires further development outside the seed before germination in the strict sense of the word (i.e. radicle emergence) is complete. In fact, de Queiroz (Reference de Queiroz1986) defined the germination in seeds of Euterpe edulis as occurring in two distinct stages: (1) protrusion of the cotyledonary petiole, and (2) growth (emergence) of the radicle and shoot from the cotyledonary petiole, which he described as ‘the germination itself’. He found that the second stage was completed about 6 weeks after the first stage. The whole-seed germination process in palms, as described by Pinheiro (Reference Pinheiro2001) for Schippia concolor, in which germination is a variation of the two remote types described below, begins with protrusion of the cotyledonary petiole and ends with emergence of the plumule through the base of the cotyledonary sheath. In S. concolor, the cotyledonary petiole emerges from the (isolated) seed 8–9 d after planting, and the root and plumule from the cotyledonary petiole about 30 and 80–90 d, respectively, after planting. In the mangrove species Nypa fruticans, germination is ‘incipiently viviparous’ (Tomlinson, Reference Tomlinson1986).

In a review of information on time to germination in palms [unit(s) of germination not specified, i.e. whole fruit, seed + endocarp (‘nut’) or seed (‘kernel’)] that included 1281 published records on 457 species, seeds in about 10% of the observations germinated within 30 d, and for the other 90% time to germination ranged from between 31 and 40 to 1941 d. Seeds in about 53% of the 1241 observations germinated between days 31 and 40 to 121–130 (Orozco-Segovia et al., Reference Orozco-Segovia, Batis, Rojas-Arechiga and Mendoza2003). The facts that seeds (diaspores) of most palms take longer than 30 d (usually much longer) to germinate and have underdeveloped embryos indicate that morphophysiological dormancy (MPD) is the major dormancy class in the family. The 10% of the diaspores that germinated in ≤ 30 d would have morphological dormancy (MD) (Baskin and Baskin, Reference Baskin and Baskin2004).

All palms have hypogeal germination since the seed is not raised above the soil; in monocots with epigeal germination, the seed is raised above the soil surface by elongation of the cotyledon (Tillich, Reference Tillich2007). Three morphological types of seedlings (germination) are commonly recognized in palms: remote-tubular, remote-ligular and adjacent-ligular (Fig. 1c). In the latter two types, the cotyledonary petiole bears a ligule, a distal projection of the leaf sheath. In remote-tubular (e.g. Phoenix) and remote-ligular (e.g. Sabal, Washingtonia) types of germination, the embryonic axis (inside the cotyledonary tube/sheath) is pushed into the soil to varying depths by downward extension (positive geotropism) of the cotyledonary petiole with an enlarged basal sheath from which the seedling develops at varying distances from the seed. The persistent radicle (primary root) breaks through the base of the cotyledonary sheath, and the plumule emerges from the cleft in the cotyledonary sheath in the remote-tubular type and from the mouth of the cotyledonary sheath (ligular extension) in the remote-ligular type. Thus, in these two types of germination the seedling is moved some distance from the seed, hence ‘remote’. In the adjacent-ligular type (e.g. Archontophoenix), the cotyledonary tube elongates very little at germination, forming a ‘button’ or mass of tissue just outside the seed, from which the radicle and shoot emerge next to the seed, hence ‘adjacent’ (Tomlinson, Reference Tomlinson1960, Reference Tomlinson1990; Corner, Reference Corner1966; Jordan, Reference Jordan1970; Moore and Uhl, Reference Moore and Uhl1973; Brown, Reference Brown1976; Fong, Reference Fong1978; Uhl and Dransfield, Reference Uhl and Dransfield1987; Henderson, Reference Henderson2006; Tahir et al., Reference Tahir, Mu'azu, Khan and Iortsuun2007). Henderson (Reference Henderson2006) has pointed out that these three germination types are not completely satisfactory because of the great variation in length attained by the petiole in developed seedlings. Some modifications of these three basic germination types have been described by Tomlinson (Reference Tomlinson1960, Reference Tomlinson1990) and Pinheiro (Reference Pinheiro2001).

The germination type is related to the internal structure of the embryo. In remote-tubular germination, the plumular-radicular axis is straight with respect to the cotyledon axis, the embryo is straight, the cotyledon elongates and the radicle is persistent. The plumular-radicular axis is oblique with respect to the cotyledon axis in the remote-ligular type, the embryo is straight, the cotyledon elongates and the radicle is persistent. In the adjacent-ligular type, the plumular-radicular axis is at an obtuse angle with respect to the cotyledon axis, the embryo is curved, the cotyledon does not elongate and the radicle is not persistent; soon after germination, the radicle is replaced by adventitious roots (Tomlinson, Reference Tomlinson1960; Henderson, Reference Henderson2006).

Thus, it might be expected that the embryo does not fit the standard definition of an underdeveloped embryo, i.e. the embryo must grow inside the seed before the radicle emerges (Baskin and Baskin, Reference Baskin and Baskin2004). However, Perez et al. (Reference Perez, Criley and Baskin2008) showed that the length of the embryo of P. remota must increase 1.6-fold to displace the operculum, i.e. for the seed to germinate. Furthermore, radicle emergence, a standard definition for germination, does not occur in palms until after the embryo has definitely grown. Of the seven genera and nine species of palms included in their compilation of the kinds of seed dormancy in seed plants, Nikolaeva et al. (Reference Nikolaeva, Rasumova, Gladkova and Danilova1985) listed them as having either morphological dormancy (б), a combination of morphological dormancy and chemical dormancy (A₁-б), a combination of morphological dormancy and mechanical dormancy (A₂-б) or a combination of morphological dormancy, mechanical dormancy and non-deep physiological dormancy (A₂-б-B₁). Nikolaeva (Reference Nikolaeva1999) stated that palms ‘.. exhibit the morphological type of dormancy’.

The rejection of palm seeds as having MD (or MPD) by some investigators (e.g. Ribeiro et al., Reference Ribeiro, Oliveira and Garcia2012a; Megalhães et al., Reference Megalhães, Lopes, Ribeiro, Sant'Anna-Santos and Oliveira2013; Nazario et al., Reference Nazario, do Nascimento Ferreira, de Lima e Borges, Genovese-Marcomíni and de Mendonça2013) is based on the fact that seeds will germinate before the hypocotyl–radicle is fully differentiated or that the embryo is fully differentiated histologically. Ribeiro et al. (Reference Ribeiro, Oliveira and Garcia2012a) stated that seeds of A. aculeata did not have morphological dormancy ‘… because the intermediary degree of differentiation [of the meristematic tissues] of the embryo does not restrict its germination…’. Nazario et al. (Reference Nazario, do Nascimento Ferreira, de Lima e Borges, Genovese-Marcomíni and de Mendonça2013) concluded that the seeds of Bactris gasipaes do not have morphological dormancy because the embryo is fully differentiated histologically. Corner (Reference Corner1966) thought that palm seeds ‘… have little power of dormancy…’, since tissue development continues in the anatomically immature embryo of mature seeds while germination is arrested. This way of thinking about MD (and MPD) differs from the traditional meaning of that (these) class(es) of seed dormancy (see above).

Thus, in the second edition of Seeds: ecology, biogeography, and evolution of dormancy and germination (Baskin and Baskin, in press) we have assigned seeds of all palms included in our survey of kinds of dormancy in the various vegetation zones on Earth to either MD or MPD. A caveat here is that decisions about assignment to either MD or MPD may be based on results of the responses of isolated seeds rather than on those of seeds within the natural germination unit, i.e. in some cases a drupe with a hard endocarp. In the literature on palm germination, the word ‘seed’ is often used for seed +endocarp (Corner, Reference Corner1966; Tomlinson, Reference Tomlinson1990). Thus, since the drupe endocarp has been shown to inhibit germination of seeds inside them, some of our assignments to MD may be a case of diaspore (fruit) MPD. In other words, the seed freed from the fruit may exhibit MD and the natural germination unit MPD. An example of this is found in P. remota, in which the isolated seeds germinated much faster than the drupes (Perez et al., Reference Perez, Criley and Baskin2008). We agree with Perez (Reference Perez2009) who concluded that most palm diaspores have MPD.

It should be pointed out, however, that isolated seeds of some palm species may exhibit MPD, i.e. those isolated from fruits with a thick stoney endocarp (Hussey, Reference Hussey1958; Broschat, Reference Broschat1998; Wood and Pritchard, Reference Wood and Pritchard2003; Ribeiro et al., Reference Ribeiro, Souza, Rodrigues, Oliveira and Garcia2011; Neves et al., Reference Neves, Ribeiro, Cunha, Pimenta, Mercadante-Simoes and Lopes2013) or from those with a thin membranous endocarp (Carpenter, Reference Carpenter1987, Reference Carpenter1988). In addition, the operculum has been shown to inhibit germination of isolated palm seeds, and thus its removal promoted germination (Carpenter et al., Reference Carpenter, Ostmark and Cornell1993; Perez et al., Reference Perez, Criley and Baskin2008; Fior et al., Reference Fior, Rodrigues, Leonhardt and Schwarz2011; Ribeiro et al., Reference Ribeiro, Souza, Rodrigues, Oliveira and Garcia2011). In A. aculeata, the micropylar endosperm (endosperm cap) part of the operculum (i.e. seed coat portion of operculum removed) was inhibitory to germination (Ribeiro et al., Reference Ribeiro, Souza, Rodrigues, Oliveira and Garcia2011).

Isolated embryos of palms have been reported to be non-dormant (Hussey, Reference Hussey1958; Ribeiro et al., Reference Ribeiro, Souza, Rodrigues, Oliveira and Garcia2011, Reference Ribeiro, Oliveira and Garcia2012a). However, since removal of the endocarp, operculum or endocarp and operculum (see references above) and gibberellic acid (GA₃) treatment (Nagao and Sakai, Reference Nagao and Sakai1979; Nagao et al., Reference Nagao, Kanegawa and Sakai1980; Roberto and Habermann, Reference Roberto and Habermann2010; Ribeiro et al., Reference Ribeiro, Souza, Rodrigues, Oliveira and Garcia2011; Neves et al., Reference Neves, Ribeiro, Cunha, Pimenta, Mercadante-Simoes and Lopes2013) promote germination, it seems likely that the embryo is not completely non-dormant, i.e. it has low growth potential and thus some degree of physiological dormancy. GA₃ would be expected to promote germination in palm seeds by increasing the growth potential of the embryo and also by weakening the resistance of the (living) micropylar endosperm covering the proximal end of the cotyledonary petiole. Abscisic acid (ABA) also would be expected to play a role in palm diaspore dormancy by inhibiting the growth potential of the embryo and by weakening of the endosperm cap (da Silva et al., Reference da Silva, Toorop, van Aelst and Hilhorst2004; Finch-Savage and Leubner-Metzger, Reference Finch-Savage and Leubner-Metzger2006). With regard to germination of oil palm seeds, Hussey (Reference Hussey1958) stated that ‘.. rupture of the operculum appears to be dependent upon intercellular breakdown in the abscission layer [of the micropylar endosperm] as well as on growth pressure of the embryo’. Jiménez et al. (Reference Jiménez, Guevara, Herrera, Alizaga and Bangerth2008) found that a sharp reduction in concentration of ABA in the embryo of oil palm was correlated with dormancy break (heat treatment). However, changes in concentration of GAs and other plant growth regulators could not be related to release of dormancy. Much remains to be learned about germination of palm diaspores, both at the whole-seed and biochemical–molecular levels.