Young lineages in an old desert. Evolution of Eulychnia

Separated sites and the question of origin

Columnar Eulychnia grow in southern Peru and Chile along one of the continent's driest zones. In the north, many sites are associated with fog-influenced coastal mountains. Farther south, some plants benefit from seasonal rainfall and also occupy sites away from the shore. Today's range comprises habitat fragments separated by extremely dry areas. This pattern raises questions about former population connections and the timing of their separation. [Merklinger et al. 2021]

In 2021, Merklinger and colleagues compared plastid data, covering selected chloroplast DNA fragments, with SNP data obtained from many genomic regions. SNPs are positions where the sequences of the plants studied differ by a single nucleotide. Combining the two data types allowed the stability of larger groups to be checked and revealed differences in the histories of individual parts of the genome. [Merklinger et al. 2021]

A DNA sample has a name and a locality

Material was collected in Chile and Peru in 2016–2019. Type localities of almost all previously described taxa were visited. The oldest available species or variety names were used on the trees for samples from these locations. The authors noted that this principle of identifying material did not constitute a taxonomic decision. A sample's former name is therefore read together with its locality, while the current atlas profile is linked to the presently accepted taxon. [Merklinger et al. 2021]

An example is Eulychnia breviflora (material labelled E. saint-pieana or E. barquitensis). POWO treats these former names as its synonyms. In turn, Eulychnia elata (E. acida var. elata in the paper) and Eulychnia chorosensis (E. acida var. procumbens in the paper) are now accepted species. The original labels document which plants were actually studied and remain necessary when comparing publications. [Merklinger et al. 2021] [2-WWW] [2-WWW] [2-WWW]

A special case concerned Camaraca near Arica. Plants could not be recollected at the site, and the authors considered the local population probably extinct. The analysis included cultivated material from Ritter's collection labelled E. aricensis. POWO now assigns this name to Eulychnia iquiquensis. Retaining provenance thus allowed inclusion of a lineage that was difficult to obtain again in the field. [Merklinger et al. 2021] [2-WWW]

Sister genera with different growth forms

Both main datasets supported close relatedness between Eulychnia and Austrocactus. These are sister genera, two lineages arising from a common ancestor in the reconstructed tree. Today's plants differ markedly in habit and environment. The branched columns of Eulychnia are associated with the western coast, and Austrocactus with southern and Patagonian habitats. Similarity of external form was not the basis for this relationship, which was established using DNA sequences. [Merklinger et al. 2021]

In the genomic analysis, Peruvian Eulychnia ritteri separated as the sister lineage to the entire Chilean group studied. The latter comprised two large groups. One included E. breviflora, E. iquiquensis and E. taltalensis. The other included E. acida, E. castanea, E. chorosensis and E. vallenarensis. Differences in flower structure, particularly hairiness and pericarpel armature, accompanied the larger groups. [Merklinger et al. 2021]

The molecular clock and its uncertainty interval

Dating of the plastid tree estimated the separation of Eulychnia and Austrocactus at about 6,70 million years ago. The highest posterior density interval, covering 95% of the estimated-age distribution and designated HPD in the paper, extended from 3,28 to 10,51 million years. For the last common ancestor of the Eulychnia lineages studied, the estimate was about 2,19 million years, with an interval of 0,88–3,84 million years. Such broad ranges describe the timescale while retaining uncertainty about the precise divergence time. [Merklinger et al. 2021]

Calibration was secondary. The authors used age estimates for larger groups from earlier family analyses because direct fossil calibrations for these cacti were unavailable. Estimates from the plastid tree were then used to date the genomic tree. The two stages therefore share some of the same assumptions. Their results are reconstructions based on a model of sequence changes and calibration, rather than two independent dates read from the fossil record. [Merklinger et al. 2021]

Wetter phases connect habitats

The authors propose a scenario in which wetter Pleistocene periods increased the continuity of suitable sites. Populations could expand their ranges and make contact again. In drier phases, they remained in separate refugia, environmental fragments allowing survival. Reduced gene exchange between these fragments would favour lineage diversification. Agreement between the estimated timescale and the history of climate changes supports this scenario, although it alone does not identify a single cause of speciation. [Merklinger et al. 2021]

Reconstruction of the area of origin suggested the importance of coastal zones. Its details depended, however, on the permitted number of regions occupied by an ancestor and on the biogeographical model used. No single certain direction of expansion from north to south or the reverse was obtained. The conclusion about the coast's role is therefore stronger than the detailed course of movement between parts of Chile and Peru. [Merklinger et al. 2021]

Different histories of plastids and the rest of the genome

Plastid and genomic phylogenies differed in the position of E. ritteri. In the first, it was sister to the Chilean E. breviflora group, and in the second, to both Chilean groups combined. Such discordance can arise after past gene flow between lineages or through incomplete lineage sorting. In the second process, DNA variants present in a common ancestor pass into descendant populations and for a time do not form a uniform history across all markers. [Merklinger et al. 2021]

Analyses also indicated possible introgression, the passage of DNA variants after hybridisation and subsequent crossing of offspring. E. castanea material from Tongoy showed a different position and a mixing signal compared with some other samples. The authors interpreted this case as a possible trace of contact with the E. breviflora lineage. The signal concerned the Tongoy samples compared. [Merklinger et al. 2021]

Species boundaries and the history of a local population

In recent diversification, population separation, accumulation of morphological differences and establishment of genomic divergence can proceed at different rates. In the study discussed, not all historical names defined uniform sample groups under every method. Some patterns suggested additional geographic structure and others broader taxon boundaries. The data show which hypotheses merit testing through denser sampling, trait observations and studies of possible population contacts. [Merklinger et al. 2021]

For E. ritteri, additional field surveys documented five populations in Caravelí and differences in their condition. Lineage distinctness in the DNA tree and the state of a particular site describe different properties of the plants. Combined information allows both the origin and contemporary biology of the taxon to be presented. [Merklinger et al. 2021] [Pauca-Tanco et al. 2021]