# Past and present genetic structure of the tropical rainforest palm Astrocaryum mexicanum: effects of anthropogenic fragmentation

## TL;DR

This study examines how forest fragmentation affects the genetic structure of a tropical palm species in Mexico, finding that genetic diversity remains high despite habitat changes.

## Contribution

The study introduces novel microsatellite loci and controls for pre-fragmentation genetic structure to assess fragmentation effects.

## Key findings

- Fragments were not genetically distinct from continuous forest populations.
- High historical connectivity implies most genetic variation exists within subpopulations.
- Bottleneck simulations suggest insufficient time since fragmentation to detect genetic variation loss.

## Abstract

To assess whether fragmentation of the lowland rainforest of Los Tuxtlas natural reserve has altered the genetic structure of understory palm Astrocaryum mexicanum, we analyzed populations from undisturbed forest and forest fragments. The questions that this study addressed were: Has habitat fragmentation reduced gene flow and within-population genetic variation (allele loss)? Has this process, in turn, increased population genetic differentiation of populations in fragments? We expected that reduced population sizes and gene flow, in fragments, has increased the effects of genetic drift, thus affecting genetic structure. The design of the study allows control for pre-fragmentation genetic structure, a common criticism against fragmentation studies, and addresses this question for a community-level important palm tree species of the tropical rain forests of southern Mexico. We sampled two cohorts (i.e., pre- and post-fragmentation palms) in each of eight populations (thirty individuals per cohort or generation, and a separation of four to nine km between populations), one composed by adult palms (80–140 years old), and the other of seedlings <3 years old. We estimated RST, inbreeding coefficient, number of alleles, heterozygosity, linkage disequilibrium, and number of migrants per generation, using variation at eight novel microsatellite loci, developed ex profeso for this study. Results indicate lack of differentiation between population pairs, and most genetic variation exists within subpopulations, implying high historical connectivity. Fragments were not genetically distinct from continuous forest populations. Simulations suggest a severe effective population size reduction at the outset of the Last Glacial Maximum 26,000 YBP, after which the area was recolonized by individuals from Central America. It is possible that the number of reproductive events that have passed since the onset of fragmentation has been insufficient to detect an effect on genetic variation, or that the extant number of palm trees in fragments is high enough to maintain the genetic diversity; bottleneck simulations agree with the first explanation. Notwithstanding, evidence suggests that populations in fragments face harsher environmental conditions, selecting against homozygotes, a situation that can jeopardize their persistence in fragments if population size is too small.