Genetic bottlenecks in time and space: reconstructing invasions from contemporary and historical collections

Abstract

Herbarium accession data offer a useful historical botanical perspective and have been used to track the spread of plant invasions through time and space. Nevertheless, few studies have utilised this resource for genetic analysis to reconstruct a more complete picture of historical invasion dynamics, including the occurrence of separate introduction events. In this study, we combined nuclear and chloroplast microsatellite analyses of contemporary and historical collections of Senecio madagascariensis, a globally invasive weed first introduced to Australia c. 1918 from its native South Africa. Analysis of nuclear microsatellites, together with temporal spread data and simulations of herbarium voucher sampling, revealed distinct introductions to south-eastern Australia and mid-eastern Australia. Genetic diversity of the south-eastern invasive population was lower than in the native range, but higher than in the mid-eastern invasion. In the invasive range, despite its low resolution, our chloroplast microsatellite data revealed the occurrence of new haplotypes over time, probably as the result of subsequent introduction(s) to Australia from the native range during the latter half of the 20th century. Our work demonstrates how molecular studies of contemporary and historical field collections can be combined to reconstruct a more complete picture of the invasion history of introduced taxa. Further, our study indicates that a survey of contemporary samples only (as undertaken for the majority of invasive species studies) would be insufficient to identify potential source populations and occurrence of multiple introductions.

Figure 1. Senecio madagascariensis sampling locations.

Geographic locations of: herbarium records sampled in Australia (A); contemporary collections in Australia (B, C) and South Africa (F). Extent maps indicating sampling areas in country-wide context (D, E).

Figure 2. Results of structure analyses.

Graphical outputs of all structure analyses undertaken; all samples (top) showing K = 3 genetic clusters; Australia only (middle) showing K = 2 clusters; and South Africa only (bottom) showing K = 2 clusters. Sampling site names are listed above their respective outputs.

Figure 3. Maps illustrating the spread of Senecio madagascariensis in Australia through time.

Density of herbarium records and location of chloroplast haplotypes (A-C); location of P1 and P2 derived from nuclear microsatellite data from contemporary field collections (as defined by clusters in the program structure [44]) (C, D); clustering of sites in Far North Queensland with P1 and P2 (D); extent of maps in relation to Australia as a whole (E).

Figure 4. Location of Senecio madagascariensis haplotypes based on three chloroplast microsatellite loci in contemporary samples from South Africa and historical samples from Australia.

The proportion of haplotypes found at each sampled site in South Africa (A); haplotypes from all herbarium records in Australia, according to their position in either the south-eastern Australian population P1 or mid-eastern Australian population P2 (B). Size of pie charts are proportional to the number of individuals sampled; median joining network of S. madagascariensis, where the smallest connector length represents one character change (C); map extents (D). Colour codes for haplotypes are consistent throughout.