Seeds Seed size

The size of seeds interests comparative ecologists because it is so variable among species. The dry mass of seeds ranges over at least six orders of magnitude across species of tropical rain-forest tree. The Melastomataceae and Rubiaceae include tropical tree species with seeds of dry mass as little as 20 g (Metcalfe & Grubb 1995; Grubb & Metcalfe 1996). At the other extreme, the seeds of a number of trees, notably legumes, approach 100 g dry mass. Within any tropical forest site, most studies have shown ranges of least five orders of magnitude for tree seed mass (Hammond & Brown 1995; Metcalfe & Grubb 1995; Grubb & Coomes 1997; Lord et al. 1997). Of course, mass is a volume-dependent property and so will rise with the cube of the linear dimensions involved, which will rapidly exaggerate size differences between species, but a million-fold range in offspring size is still enormous when compared with animal groups.

Seed size might be under allometric control of other characters. There is evidence of correlations with other size variables. The difficulties of small plants producing big seeds and of small fruits containing big seeds will probably always lead to some degree of positive correlation between plant size and seed size (Fig. 5.1). For tropical trees, a number of studies have shown increases in seed size with adult stature within a particular forest (Hilty 1980; Foster & Janson 1985; Metcalfe & Grubb 1995; Hammond & Brown 1995; Kelly 1995; Grubb & Coomes 1997). Dispersal mechanism may also influence seed size (Fig. 5.1), possibly because of disperser selection for fruit size. In three neotropical sites there was a consistent pattern of mammal-dispersed seeds being 12-14 times heavier on average than bird-dispersed seeds (Hammond & Brown 1995). This might be due to mammalian frugivores choosing the big end of the fruit-size range available in the forest, making it appear as though there was some evolutionary relationship. However, Kelly (1995) found that this pattern was often repeated within different genera at a site in Peru, indicating recurrent evolution of large seed size in mammal-dispersed species.

A feature of tropical rain forests is the presence of very large-seeded species

Figure 5.1 Seed mass versus tree height for species from Manu, Peru. Species categorised by dispersal mode (circles, mammal-dispersed; triangles, bird-dispersed) and successional status (filled, non-pioneers; open, pioneers). Data from Foster & Janson (1985).

(Lord et al. 1997). Foster (1986), in a review of the adaptive value of large seeds, proposed several advantages. However, critical analysis does not support the acceptance of all of them. Foster put forward the following as likely benefits for bigger seeds.

Improved seed longevity

Simplistically, one might argue that bigger seeds will have more reserves and therefore will be expected to survive longer. Foster (1986) pointed out a significant positive correlation between time to germination and seed size in Ng's large data set for Malaysian tree species (Ng 1980). However, Hopkins & Graham (1987) found that it was mostly small seeds that survived burial of up to 2 years well in Queensland (Fig. 5.2). Kanzaki et al. (1997) also found a significant negative correlation between seed size and survival time in tropical forest soil for tree species from Malaysia. Respiration rate measurements on some tropical seeds shed light on this paradox. Garwood & Lighton (1990) discovered that seed water content was of greater influence on oxygen consumption rate than seed size. Dormant seeds have low water contents and, hence, low respiration rates. Large seeds tend to have high water contents and therefore respire rapidly. For seeds of similar water content, large seeds have higher absolute respiration rates, but frequently lower rates per unit dry mass. In theory, large seeds that could dry out to enter dormancy should be able to survive for long periods, but this appears to be a relatively rare strategy. Large dormant seeds might be particularly susceptible to vertebrate seed predators in the forest.

Greater room for secondary compounds

A large seed might be able to store more chemical defences. These could possibly be mobilised to defend the seed against attack by pathogens or small invertebrate attackers. However, for large seed-eaters the total concentration of chemical defences is more likely to be influential on food choice. It is possible, though, that very large seeds might contain more than the safe

Figure 5.2 Viability of seeds against time buried in soil for species from Queensland, Australia. Solid symbols indicate primary forest species (mostly large seeds), hollow symbols, secondary forest species (mostly small seeds). Data from Hopkins & Graham (1987).

10"4 10"2 1 100

Figure 5.3 LAR as a function of seed dry mass in small seedlings of 43 tropical tree species. Data from ter Steege (1994a), Osunkoya et al. (1994) and Kitajima (1994).

10"4 10"2 1 100

Seed dry mass (g)

Figure 5.3 LAR as a function of seed dry mass in small seedlings of 43 tropical tree species. Data from ter Steege (1994a), Osunkoya et al. (1994) and Kitajima (1994).

dose of a toxin and only part of the seed would be eaten by any one animal at one time, improving the chances of the embryo surviving.

Increasedphotosynthetic to non-photosynthetic tissue ratio Large seed reserves might allow faster development of greater photo-synthetic areas in newly germinated seedlings and hence a greater chance of survival in the shade. This does not seem to be the case. Kitajima (1992) found that there was a strong negative correlation between seed mass and relative photosynthetic ability of newly germinated seedlings of tropical tree species. Small seeds tend to produce seedlings with thin cotyledons of relatively large area (Fig. 5.3), allowing small seedlings to achieve substantially greater relative growth rates. Tiny tropical tree seedlings are capable of persisting in very deep shade (Ellison et al. 1993; Metcalfe & Grubb 1997).

Growth into higher light or deeper soil

Bigger seeds produce bigger seedlings (Janos 1980; Howe & Richter 1982), and there can be an advantage in greater seedling size. There may be strong gradients in light availability above the forest floor and in water availability beneath it. Large seeds can produce taller seedlings that get more light, or deeper-rooted seedlings that get more water. The seedlings from big seeds may germinate successfully from under greater depths of soil or litter, and can probably compete with neighbours more vigorously and develop mycorrhizas more successfully (Janos 1980). Greater reserves will allow greater chances of survival during unfavourable intervals e.g. when shaded by a fallen leaf, or during a cloudy period. Boot (1996) found a positive correlation between survival of germinated seedlings in the dark over one year and seed size for six tree species from Guyana. Germinants of large-seeded species (e.g. Castanospermum australe) in Queensland, Australia, could survive burial for two years (Hopkins & Graham 1987).

Tolerance of damage or tissue loss

Damage due to pathogens, herbivores and falling debris is likely to affect seed and seedling survival in the forest. The greater reserves of large seeds can help cope with these vicissitudes. Two very large-seeded species (more than 100 g fresh mass) from New Guinea were more tolerant, in terms of seedling survival and height growth relative to control plants, of removal of more than 50% of non-embryo tissue from the seeds than two other species of seed size more than an order of magnitude smaller (Mack 1998). Fragments of large seeds may also be capable of developing into plantlets as has been shown for Gustavia superba (Harms et al. 1997). Harms & Dalling (1997) measured the ability of seedlings to recover from decapitation for 13 species with seed fresh mass in the range 0.2-107.6 g. Only the five species with seed mass of 5 g or greater survived decapitation and resprouted. They all had hypogeal germination.

The evidence points to large seed size being advantageous because it produces larger seedlings with reserves sufficient to meet the requirements of physical and chemical defence, periods of resource shortage and repair of damage (Kitajima 1996). The abundance of tree species with very large seeds in tropical rain forests may reflect the deep shade of the forest understorey with fierce competition and ever-present threats from pests and diseases. The presence of potential seed dispersers of large body size may also favour large seeds in the tropical rain forest.

However, many species still have small seeds, probably because of the basic evolutionary tradeoff between seed size and seed number. If a tree has a particular amount of resources it can devote to reproduction then to produce more seeds it will have to reduce the amount of resources given to each seed. More seeds will mean more attempts to produce viable descendants in the next generation. But generally reduced size means increased risk of mortality. So natural selection balances the increasing number of attempts against the reducing chance of survival and comes up with a seed size appropriate to the species. In essence, small seeds are riskier but more can be produced per unit investment.

The tradeoff of seed size against seed number may not be as direct as the foregoing paragraph made it seem, at least as far as fleshy fruits are concerned. Fruits with small seeds tend to have a higher ratio of flesh dry mass to seed dry mass (Grubb 1998b). This is probably because small fruits require to offer proportionally larger rewards than big fruits to attract dispersers.

There may be situations where seed size is less relevant to survival, or where small seeds gain special advantages. It has been argued that in the relatively resource-rich environment of gaps seed size may be less influential on seedling survival and that seedlings from small seeds may rapidly catch up with those from larger ones because of their higher relative growth rates. The idea that strongly light-demanding species have smaller seeds than shade-tolerant ones is quite firmly entrenched in the ecological literature (Swaine & Whitmore 1988), but it has been challenged of late (Kelly & Purvis 1993; Grubb & Metcalfe 1996). Comparisons of seed size between species believed to be gap-demanding for regeneration and those that can persist as juveniles in the forest understorey have generally shown greater mean seed mass (or other size measure) for shade-tolerants in tropical rain forests (Foster & Janson 1985; Hammond & Brown 1995), but this has always involved a very large range of seed size in both species groups. However, statistical analyses that correct for possible phylogenetic influences have shown much less clear evidence for larger seed size in shade-tolerants (Kelly & Purvis 1993; Grubb & Metcalfe 1996). Grubb & Metcalfe (1996) argued that these analyses, which imply a strong phylogenetic influence on seed size, indicate that this character exhibits evolutionary inertia. Within a particular clade, seed size changes slowly in evolutionary terms, and it may not be particularly influential as to whether a species can or cannot be successful as a shade-tolerant or gap-demanding tree. Large-seeded light-demanders include Aleurites moluccana (7.8 g dry mass) (Grubb & Metcalfe 1996) and Ricinodendron heudelotii (1.4g dry mass) (Kyereh et al. 1999). Various authors (Metcalfe & Grubb 1995; Grubb 1996; Grubb & Metcalfe 1996; Hammond & Brown 1995) have highlighted the presence of very small-seeded, strongly shade-tolerant species among rain-forest tree floras. Hammond & Brown (1995) hypothesised that small adult stature, low resource availability in the forest understorey and use of small-bodied dispersers compound to favour small seed size. These are likely to be selection pressures acting in favour of a trend of decreasing seed size with distance below the top of the canopy, but they are probably still insufficient to explain the very tiny seeds of some species, notably among the Rubiaceae and Melastomataceae. These minuscule seeds are possibly specialist exploiters of certain regeneration sites in the forest. Given the small size of the seeds and their subsequent seedlings, regeneration could not be successful if litter were present, so litter-free microsites on the forest floor are required. Sites on steep banks and slopes are most suitable, where large seeds would fall down, but the minute seeds are caught by tiny irregularities in the soil surface (Grubb & Metcalfe 1996; Metcalfe et al. 1998). Grubb & Metcalfe (1996) point to the possibility that these species have evolved from light-demanding ancestors.

Tall trees in the caatinga of Venezuela were shown to have smaller seeds than species of similar stature from forests on more fertile soils nearby (Grubb & Coomes 1997). This was interpreted as the caatinga species maintaining seed number in an environment where soil nutrient poverty had reduced the resources available for reproduction. The same authors argued that root competition was more intense on such low fertility sites (Coomes & Grubb 1998a) which one might predict would favour larger-seeded species, but the possible benefits of increased seed size rise too slowly to outweigh the disadvantages of fewer seeds produced.

Was this article helpful?

0 0

Post a comment