Tests of the escape hypothesis

There have been many studies of the survival of seeds and seedlings of tropical trees with respect to conspecific density or distance to adult trees (Hammond & Brown 1998) (Table 4.10). Some of the research published had poor degrees of replication, studying just the offspring of one tree at one fruiting season. Despite this, the evidence seems to be in favour of the escape hypothesis. Seedlings of Dipteryx panamensis at La Selva, Costa Rica, showed a 100% mortality over the 7-20 month old period within 8 m of the parent tree (Clark & Clark 1984). The seeds of Virola nobilis are attacked by weevils (Curculionidae) and bark beetles (Nitidulidae) and probably mammals (Howe 1993). The mortality due to these predators was very high near parent trees. There was a 22-44-fold advantage in survival for seeds getting 45 m from the parent tree compared with 5 m away (Fig. 4.11). This appeared to be mostly a density-dependent effect because seeds 5 m from an adult male tree had the same survival rate as those 45 m from a fruiting female tree. Well-dispersed seeds may escape death from insect attack because the pests take longer to find them giving the seeds an opportunity window in which to germinate and establish. Chlorocardium rodiei seeds in a Guyanan forest were equally likely to be infested by beetles over an 85 week period at a range of dipseral distances, but those dispersing further were more likely to establish as the beetle attack came later (Hammond et al. 1999).

Besides animals, other agents of mortality have been shown to act in a density- and distance-dependent manner. Augspurger (1983a) found that seedling mortality due to fungal pathogens was greater closer to parent trees of Platypodium elegans. A canker disease of Ocotea whitei that led to sapling mortality on Barro Colorado Island also exhibited density dependence (Gilbert et al. 1994). Yamada & Suzuki (1997) identified parental leaf litter as the factor inhibiting seedling establishment beneath seeding trees of Scaphium

Table 4.10. Comparison between vertebrate and insect attack on seeds and seedlings of tropical rain-forest tree species based on their support of the escape hypothesis n, The number of mature individuals around which seed/seedling survival was studied; h/l indicates a comparison between high- and low-density sites for mature individuals; nd, no data available to author.

Seeds or

Species

n

Attack type

seedlings or both?

Supports escape hypothesis?

Aglaia mackiana

6

insect/pathogen

seedlings

no

Astrocaryum macrocalyx

2

insect

seeds

yes

Carapa guianensis

nd

insect

seeds

yes

Chlorocardium rodiei

10

insect

seeds

yes

Copaifera pubiflora

20

insect

seeds

yes

Gilbertiodendron dewevrei

h/l

insect

seeds

yes

Julbernardia seretii

h/l

insect

seeds

yes

Mora gonggrijpii

nd

insect

seeds

no

Macoubea guianensis

4

insect

seeds

no

Maximiliana maripa

6

insect

seeds

yes

Normanbya normanbyi9

5

insect

seeds

yes

Pouteria sp.

4

insect

seeds

no

Scheelea zonensis 0

7

insect

seeds

yes

Scheelea zonensis11

14

insect

seeds

yes

Virola nobilis

5

insect

seeds

yes

Virola michelii

nd

insect

seeds

no

Astrocaryum macrocalyx

2

vertebrate

seeds

no

Bertholletia excelsa

4

vertebrate

seeds

no

Brosimum alicastrum

14

vertebrate

seeds

no

Carapa guianensis

nd

vertebrate

seeds

no

Chlorocardium rodiei

10

vertebrate

both

no

Dipterocarpus acutangulus

1

vertebrate

seedlings

no

Dipterocarpus globosus

1

vertebrate

seedlings

no

Dipteryx micrantha

?

vertebrate

seeds

no

Dipteryx panamensis

6

vertebrate

seedlings

yes

Dipteryx panamensis

19

vertebrate

seeds

no?

Dipteryx panamensis

nd

vertebrate

seeds

no

Dryobalanops aromatica

1

vertebrate

both

no

Dryobalanops lanceolata

1

vertebrate

both

no

Eperua grandiflora

1

vertebrate

both

no

Eperua falcata 9

2

vertebrate

both

no

Gilbertiodendron dewevrei

h/l

vertebrate

both

no

Gustavia superba 0

h/l

vertebrate

seed

no?

Gustavia superba

h/l

vertebrate

seed

no

Hymenaea courbaril

2

vertebrate

seed

no

Julbernardia seretii

h/l

vertebrate

both

no

Macoubea guianensis

4

vertebrate

seeds

no

Normanbya normanbyi9

5

vertebrate

seeds

no

Pouteria sp.

4

vertebrate

seeds

no

Scheelea zonensis

28

vertebrate

seeds

yes

Tachigali versicolor

2

vertebrate

both

no

Table 4.10. (cont.)

Seeds or

seedlings or

Supports escape

Species

n Attack type

both?

hypothesis?

Virola michelii

nd vertebrate

seeds

no

Virola nobilis

5 vertebrate

seeds

no

Density, but not distance, effect detected.

After Hammond & Brown (1998). Individual studies: Mack et al. (1999); Terborgh et al. (1993); Hammond & Brown (1998); Hammond et al. (1999); Ramirez & Arroyo (1987); Hart (1995); Notman etal. (1996); Fragoso (1997); 9Lott etal. (1995); °Wright (1983);

Wilson &Janzen (1972); Howe etal. (1985); Burkey (1994); Itoh etal. (1995); Clark & Clark (1984); De Steven & Putz (1984); Forget (1993); Forget (1992a); 9Forget (1989); °Sork (1987); Forget (1992b); Forget etal. (1994); Kitajima & Augspurger (1989).

Density, but not distance, effect detected.

After Hammond & Brown (1998). Individual studies: Mack et al. (1999); Terborgh et al. (1993); Hammond & Brown (1998); Hammond et al. (1999); Ramirez & Arroyo (1987); Hart (1995); Notman etal. (1996); Fragoso (1997); 9Lott etal. (1995); °Wright (1983);

Wilson &Janzen (1972); Howe etal. (1985); Burkey (1994); Itoh etal. (1995); Clark & Clark (1984); De Steven & Putz (1984); Forget (1993); Forget (1992a); 9Forget (1989); °Sork (1987); Forget (1992b); Forget etal. (1994); Kitajima & Augspurger (1989).

macropodum. The thick layer of large leaves appeared to prevent root penetration to the soil in the germinating seeds.

After reviewing the evidence from the literature (Table 4.8), Hammond & Brown (1998) came to the conclusion that density- and distance-dependent mortality was generally exhibited more strongly where invertebrates were the main agents of destruction rather than vertebrates. This, they argued, reflected the different food choices and foraging strategies of the two groups. Seed-eating insects are frequently highly specialised, attacking a very limited range of species and foraging based on cues such as chemical attractants that are likely to lead them to sites of high seed density. Vertebrates are more catholic in food choice and will forage more opportunistically. They also seem easier to satiate than invertebrate predators that increase rapidly in population size in response to large seed crops.

On Barro Colorado Island, Schupp (1992) found that Faramea occidentalis showed clear density-dependent seed mortality around parent trees, but when the forest was considered as patches of low and high adult Faramea population densities, seed survival was higher in the areas of forest with more adult Faramea trees. This Schupp ascribed to satiation of territorial rodents in the areas of the forest with denser Faramea populations. Burkey (1994) found that seed predation rates by small mammals near individuals of Brosimum alicastrum at Los Tuxtlas, Mexico, were inversely related to tree seed-crop size (Fig. 4.12). Those trees producing big seed crops were more effective at satiating the rodents.

Augspurger & Kitajima (1992) manipulated the seed distribution of two individuals of Tachigali versicolor on Barro Colorado Island, Panama. They created transects of even distribution of seeds, extending the tail of the distribution beyond the normal maximal seed dispersal distance to 100— 1800 m. Mortality was largely due to mammalian predation of seeds and early-stage seedlings. Density-dependent mortality was found to act at two spatial scales. At the fine scale there was higher mortality among the dense populations of seeds and seedlings near the parent tree. On a broader scale, high mortality occurred in the extended tail of the seed distribution. Augspurger & Kitajima argued that this reflected predator satiation near the parent. The distribution of established seedlings after mortality was still strongly skewed toward the parent tree. This might not be disadvantageous in this species, because, being monocarpic, the death of the seed parent would leave a gap that could be filled by one of its own offspring that fell beneath it.

In the Ituri Forest of Congo, Hart (1995) studied the seed and seedling survival of two caesalpinoid legume trees with very big seeds. Gilbertioden-dron dewevrei forms mbau forest groves where it is the single dominant species. Julbernardia seretii occurs in the more species-rich matrix forest around the mbau patches, but it also tends to be clumped in adult distribution. Both species possess relatively ineffective ballistochorous seed dispersal, with Gilbertiodendron getting few seeds more than 10 m from the parent crown, and Julbernardia doing slightly better with maximal dispersal distan-

Figure 4.11 Seedling survival to 12 weeks after fruit fall as a function of distance from 17 fruiting Virola nobilis trees for 3400 seeds planted in 1982. Twelve weeks marks the end of dependence on parental endosperm, and of vulnerability to insect seed predators. After Howe (1993).
Figure 4.12 Seed prédation as a function of seed production in trees of Brosimum alicastrum growing at Los Tuxtlas, Mexico. After Burkey (1994).

ces of 30-40 m. Hart found that seed predation of Gilbertiodendron was very high, with only 0.3% seed survival. Seed predators included two curculionid beetles that did not appear to exhibit satiation, and mammals (antelopes and rodents) which probably were satiated in the groves. The mammalian predators were efficient destroyers of any seeds that were dispersed out of the mbau grove. Seed survival was best at the periphery of the mbau grove. Julbernardia showed a much higher seed survival, about 60%. Gilbertiodendron had a high seedling survival, 49% over 10 years, allowing it to maintain its dominance in the mbau groves.

Another grove-forming tropical tree species is the brazil-nut tree (Berthol-letia excelsa) of Amazonia. Peres et al. (1997) demonstrated that predator satiation did not occur in this species, quite the reverse. The rate of seed removal from the large pyxidia produced by the brazil-nut trees was higher in the Bertholletia groves than outside them. Agoutis were the main removers of the brazil nuts. Peres et al. argued that the agoutis outside the brazil-nut groves did not have the search image for the brazil nuts and so ignored them. Brazil-nut removal increased in the groves at the period of maximum pyxidia production. Thus there was no evidence of predator satiation in this system, which is beneficial to the brazil-nut trees because the agoutis are the main dispersers of the species through scatter-hoarding.

Many studies equate seed removal by mammals with predation, but if scatter-hoarders are responsible a proportion of the removed seeds may survive (Forget et al. 1998). Burial of seeds by caching mammals may be important in some species because this allows the seeds to escape insect attack (Terborgh etal. 1993).

In conclusion, there is evidence that the escape hypothesis is correct and there is an advantage in terms of survival to dispersal to greater distances from the parent tree, particularly where insects are the main predators. Vertebrate predation has rarely been documented to operate on tropical trees in a manner that makes greater dispersal distance strongly advantageous in terms of survival. Satiation of vertebrate predators may blur some of the expected patterns of offspring survival. Invertebrates seem less liable to satiation.

Many studies of seed and seedling survival have been conducted for large-seeded species because of the logistic advantages of working with readily visible objects in relatively low numbers. The fates of the tiny seeds of many species of tropical tree have been much less well studied. It seems less likely that predator satiation will operate with very small seeds. Another factor that is likely to disrupt the recruitment pattern predicted from knowledge of the effectiveness of seed dispersal and predation is the presence of superior sites for establishment in the forest. This is dealt with in the next section.

Was this article helpful?

0 0

Post a comment