Stirring the Gene Pool: Will Data on Trees Lead to Forest of Understanding?

Understanding the genetic structure of ecosystems is important in saving biodiversity. That's especially vital as humans fragment vast natural areas into isolated patches, cutting off avenues for outcrossing. Unfortunately, scientists' understanding of breeding systems and gene flow may be even more fragmented than the ecosystems. Botanists are hard at work remedying the situation. Using genetic markers, they analyze plant paternity and infer how far pollen travels on gentle breezes or the backs and bills of animal pollinators. By comparison, however, not much is known about maternal gene flow in seeds after fertilization. "Seed movement is tougher to analyze, so it's been less well studied," says James Hamrick, population biologist at the University of Georgia. Matthew Hamilton, assistant professor of biology at Georgetown University, has now partially filled the seed void using the tropical tree Corythophora alta. As major primary producers in terrestrial ecosystems, trees provide food and protection for a host of other species. Hamilton sampled 162 mature trees (trunks at least 10 cm in diameter at breast height) spread across seven populations in 1-10 hectare plots of fragmented and continuous Amazonian forest.1 Reasoning that chloroplasts containing their own genes are transmitted exclusively in the maternal line in most plants, Hamilton assessed previous seed dispersal patterns by assaying trees for chloroplast DNA markers, or polymorphisms. Hamilton expected wide seed dispersal, but that's not the case. The seven populations consisted of just two chloroplast haplotypes--combinations of DNA markers--with each population having just one haplotype. In other words, Corythophora populations show surprisingly little genetic variation when it comes to chloroplasts, and therefore very little seed mobility. "No one has ever seen such localized population structure for chloroplast DNA," says Hamilton. Hamrick thinks Hamilton's data are "exceptional" in showing only two haplotypes. But what accounts for the pattern? "Because he's only seeing two haplotypes out of many possibilities, it tells me there's some history there." In other words, do the data reflect past patterns, before fragmentation? Adds Hamrick, "he might have found the exact same genetic structure if he'd gone in before fragmentation." That's precisely what Hamilton thinks: "The majority of trees were present and growing before the forest was cleared." The pattern was there in the continuous forest before disturbance, and "those fragments are capturing a single lineage." Hamilton is now collecting samples 50-200 km distant to find out "if this is just a local effect or a much broader pattern." But what could cause chloroplast homogeneity in a large forest? Hamrick wonders if it "could have been caused by two original parents carrying either haplotype that colonized this region." After all, the forest isn't older than about 10,000 years (going back to the last glacial period). Pre-Columbian Indians could have started the pattern by clearing and burning. Hamilton isn't convinced: "There's no evidence I'm aware of that the forest was extensively cleared." Instead, he thinks the limited variation may be due to "isolation by distance"--seeds can't move very far from the parent tree. Though tropical forests are rich in biodiversity, individuals of each species are widely spaced. Hamilton thinks some sort of local selection effect is also possible. And he hasn't ruled humans out of the equation: "Chances are that [continued] fragmentation would reduce seed dispersal even further." Hamilton plans to examine younger trees. "If fragmentation is having an effect, we should see a difference between present and past population structure." Hamilton's data seem to agree with a proposal made by Paul Ehrlich and Peter Raven in 1969 that "gene flow in nature is much more restricted than commonly thought" and "at least in many cases is of little or no importance in maintaining ... species."2 Instead, local selection pressures take precedence. That view dominated population genetics for years, but evidence to the contrary started accumulating in the 1980s. We now know that paternal genes in pollen often flow across large distances. According to Hamrick, "50-60 percent of pollen comes from outside the stand in wind-pollinated species." Now biologists are seeing the same thing for insect-pollinated tropical trees. "Some of the larger native bees will forage over several kilometers." Hamrick's own work on tropical strangler figs bears this out.3 Tropical figs are keystone species--they reproduce year-round and thereby support a wide variety of fruit-eating animals. John Nason and E. Allen Herre, together with Hamrick, used paternity analysis and a handy plant/insect mutualism to estimate the number of trees composing an interbreeding unit in several fig species--that is, the number of trees in the male phase of reproduction surrounding a maternal tree. (Female parts develop before male stamens in these flowers.) The researchers took advantage of the fact that strangler figs are serviced by female wasps that become trapped after delivering pollen, and their own eggs, to female-stage flowers. By selecting a fruit with only one trapped wasp, the researchers knew that it received pollen from a single male flower. (Wasp eggs laid at the female stage develop into adults that emerge at the male stage, carrying pollen with them.) Using protein markers called allozymes to judge genetic diversity, the group inferred breeding units of 150 to more than 750 trees--a lot more than are on the 15 km2 study site at Panama's Barro Colorado Island. In fact, breeding units may cover more than 600 km2. That kind of wide pollen base probably lets sparse fig trees reproduce in forest fragments. Unfortunately, pollination in most tropical trees doesn't involve the same kind of plant/wasp mutual admiration society, so they probably don't form large, interbreeding "metapopulations." If pollen travels so far in the tropics, why is seed dispersal so pokey, at least in Hamilton's trees? Part of the explanation may depend on who's eating the fruits and seeds. "I don't really know for sure what the dispersers are," admits Hamilton. So far, there's evidence for birds, rodents, and even primates. Hamilton also wants to look at nuclear genes and do paternity analysis to compare paternal and maternal gene flow. Resolving the issue is important. If seed dispersal is slow in other trees, how did they spread so rapidly after the last glaciation? Hamrick points to oaks: Did they migrate from the Mediterranean region to the rest of Europe in a coherent wave, or through long-distance jumps that established outpost colonies that gradually coalesced? With additional work on temperate as well as tropical species, maybe fragments of information will coalesce into a whole forest of understanding. Barry A. Palevitz (palevitz@dogwood.botany.uga.edu) is a contributing editor for The Scientist. M.B. Hamilton, "Tropical tree gene flow and seed dispersal," Nature, 401:129-30, Sept. 9, 1999. P.R. Ehrlich, P.H. Raven, "Differentiation of populations," Science, 165:1228-32, 1969. J.D. Nason et al., "The breeding structure of a tropical keystone plant resource," Nature, 391:685-7, Feb. 12, 1998.

Stirring the Gene Pool: Will Data on Trees Lead to Forest of Understanding?

Interested in reading more?

Become a Member of