Seed size of woody plants in relation to disturbance, dispersal, soil type in wet neotropical forests

Ecology, Dec, 1995 by D.S. Hammond, V.K. Brown

INTRODUCTION

Most plants suffer a higher risk of mortality as a seed or seedling than at any other stage in their lives. It is also the period during which an individual is most likely to be exposed to variable and/or adverse conditions. The combination of a high risk of mortality with a largely unpredictable set of establishment conditions places considerable importance on those adaptations that can (1) increase the likelihood of encountering favorable establishment conditions (sensu Howe and Smallwood 1982) and (2) decrease the risk of mortality upon arrival at a particular site (Foster 1986). To account for these selective pressures, a compromise between seed size and number is thought to take place (Smith and Fretwell 1974).

Seed size is the result of selective pressures being directly or indirectly applied through a variety of pathways. Direct pressures include those factors that are thought to constrain seed and seedling survival continuously over long-term ecological time, e.g., light (Turner 1990), moisture (Howe 1990) and nutrient availability (Foster 1986), dispersal (Howe et al. 1985), predation (Janzen 1970), herbivory (Coley 1980), and pathogen attack (Augspurger 1984). Indirect pressures are more the result of long-standing evolutionary pathways and influence seed size via the limitations of parental size (Foster and Janson 1985), form (Leishman and Westoby 1994), and lineage (Mazer 1990).

Life-history characteristics (viz here seed size) can respond to many different direct and indirect selective pressures (sensu Harvey and Pagel 1991). These pressures place conflicting demands on seed size, limiting the extent to which adaptation can address any single factor. In tropical forests, microclimatic, soil, and biotic factors can vary considerably. Several hypotheses have been considered that might account for seed size selection in tropical forests (Foster 1982). Three of these, viz: (1) dispersal mode, (2) light conditions (i.e., gap type), and (3) soil conditions either actively select for or passively constrain seed size, often via the subsequent advantage each of these may provide in seedling survival. The advantages associated with escaping heavy mortality near parent trees of most tropical species places a premium on seed and fruit characteristics that enhance dispersal. Seed characteristics that increase the chance of seedling establishment at the depositional site under various soil and light conditions would also be advanced during the selection process.

A wide range of light and moisture conditions develop through the gap-phase regeneration cycle. A diverse array of vertebrate and invertebrate animals disperse and/or consume seeds and seedlings. Soil conditions vary widely from permanently inundated swamp through to exceptionally well-drained upland plateaus and thin-horizoned, mountainous slopes.

In this paper, we examine to what extent gap-size preference, as required for establishment to reproductive age, dispersal mode, and soil type can explain seed size variation in a set of woody species from a tropical forest in Guyana. In addition, we analyze data available from Manu, Peru and Burro Colorado Island (BCI), Panama to compare the influence of dispersal modes and gap preferences on seed size among geographically separate sites.

METHODS

Mabura Hill data

Seeds were collected throughout the northern sections of a 600 000-ha timber concession at Mabura Hill (5 [degrees] 13[minutes] N, 58 [degrees] 48[minutes] W), which is situated between the Demerara and Essequibo rivers in central Guyana. Mabura Hill is [approximately equal to]100 m above sea level. The region has a mean annual rainfall of 2400-3000 mm and a mean annual temperature of 25 [degrees] C. The movement of the Intertropical Convergence Zone produces two relatively dry seasons extending from January-April and September-October, though monthly rainfall rarely drops to [less than]50 mm.

Fresh seeds were collected for 187 native, woody plant species from April 1992-December 1993 (see Appendix). A seed was considered to consist of an embryo, reserve (endosperm or cotyledon), and seed coat. A mature seed was considered one that could germinate and produce normal seedling structures. To test the viability of suspected immature seeds, samples were sown. Collected material was identified by staff at Mabura Hill using the guides by Roosmalen (1985) and Polak (1992), or at the Institute of Systematic Botany, University of Utrecht, The Netherlands. Specimens cover 47 families and 118 genera. Twenty-two species (11%) have yet to be placed at the generic level.

Seeds were collected from as many parental sources as possible: 36% of all species were from one source, 18% from two sources, and 46% from three sources. Twenty seeds were obtained for each species, except for 20% of the species for which fewer seeds were available (always [greater than or equal to] five).

The fresh and dry mass of seeds were measured to the nearest milligram, except for a few, very small-seeded species, which had to be weighed to the nearest microgram using more accurate balances in the UK. Dry masses were obtained by drying to constant mass at 75 [degrees] C. Fresh seed mass was highly correlated with other measures, such as dry seed mass (gram) (r = 0.91, F = 740.42, df = 1, 153, P [less than] 0.001) and seed volume (cubic centimetre) (r = 0.95, F = 1667.1, df = 1, 176, P [less than] 0.001).