Hidden diversity in Eriosyce. DNA, geography and pollinators in species recognition

Similar plants with distinct histories

A globose stem divided into ribs, dense spines and funnel-shaped flowers form the familiar appearance of many Chilean Eriosyce. Plants in successive valleys and coastal hills differ, yet the significance of those differences can be difficult to assess. Flower colour, spine thickness and body size may vary within a population. A taxonomist must establish which combinations of characters correspond to distinct evolutionary lineages. In the former E. curvispina complex, an answer emerged by combining DNA evidence, morphological observations and accurate geographical information about localities [Guerrero et al. 2019] [Walter et al. 2024].

Species boundaries affect the interpretation of all subsequent evidence. A broadly applied name may combine many localities and suggest an extensive range. Once separate lineages are recognized, each occupies only part of that area. The scope of information about flowers, seeds and pollinators attributable to a species also changes. A taxonomic revision therefore connects nomenclatural history with the biology of actual populations. Its consequences extend from identifying plants in a collection to selecting habitats that require protection [Guerrero et al. 2019] [Walter et al. 2024].

A relationship tree for the genus

A phylogeny published in 2019 provided the starting point. Researchers included 63 Eriosyce taxa representing species and infraspecific units, together with 19 outgroup representatives. The latter placed the focal plants within a wider tree of relationships. Four plastid DNA regions and one nuclear region were compared. Branching patterns were reconstructed from sequence similarities and differences, and their support was assessed. This approach supplied an independent reference for earlier classifications based primarily on appearance [Guerrero et al. 2019].

Seven principal clades were recognized within the broadly defined genus, each comprising a common ancestor and its descendants. Some branches received strong support, while one principal clade was less well supported. Certain earlier morphological groups occupied scattered positions. Similar underground organs, small stems or hairy fruits may have arisen independently in comparable environments. This finding fits the wider evolutionary history of cacti, in which adaptations to drought repeatedly produced similar forms [Guerrero et al. 2019].

The analysis also exposed difficulties with broad treatments of several species. Taxa previously included in E. curvispina occupied different positions on the tree. Denser population sampling and comparisons with plants from areas associated with the earliest names became necessary. This was a stage of formulating and testing hypotheses about species limits. Five DNA regions supplied a particular portion of the hereditary information, and a fuller reconstruction required further evidence and careful interpretation of less clearly resolved branches [Guerrero et al. 2019] [Walter et al. 2024].

Microsatellites reveal population structure

The 2022 study added a population perspective to the phylogeny. Researchers developed 12 microsatellite markers, short repeated DNA segments whose variants can be compared among individuals. Material was collected from 150 plants in nine populations, including Ocoa, Farellones, Putaendo and Los Molles. Microsatellites described the distribution of genetic variation at a much finer scale than the broad generic tree. Localities differed in their combinations of variants and degree of distinctness [Villalobos-Barrantes et al. 2022].

The two methods addressed related but different questions. Sequences from selected regions helped reconstruct relationships among lineages, whereas microsatellites described variation among the sampled populations. For Putaendo and Escorial, stronger phylogenetic separation accompanied weaker microsatellite differentiation. The authors considered contemporary gene exchange as one explanation. That inference remains an interpretation of marker patterns. Direct tracking of pollen movement, seedling parentage and gene flow across generations would provide evidence at another temporal scale [Villalobos-Barrantes et al. 2022].

Ocoa stood out for its low genetic variation relative to the other localities. Only nine plants were sampled there, limiting the precision with which the entire population could be described. The subsequent revision assigned this population to E. robusta. Combining these findings identifies a concrete subject for further research, particularly in a fragmented landscape. Both the distinctness of a lineage and the variation retained within it can matter for conservation and future generations [Villalobos-Barrantes et al. 2022] [Walter et al. 2024].

Nine species and the importance of historical names

The 2024 revision expanded sampling to additional populations, including plants from Cerro La Leona in the Cachapoal region. This area is associated with the original description of E. curvispina. Four plastid markers and one nuclear marker were again used, with results compared against morphology and nomenclatural history. Nine species were recognized, E. aconcaguensis, E. andicola, E. choapensis, E. curvispina, E. grandiflora, E. horrida, E. mutabilis, E. orientalis and E. robusta [Walter et al. 2024].

A major part of the work established how names should be applied. A botanical name is anchored to a nomenclatural type, the element to which its formal application refers. Farellones plants, previously treated broadly as E. curvispina, matched the historical Pyrrhocactus grandiflorus from Cerro Ramón in morphology and geography. The combination E. grandiflora was therefore adopted. E. curvispina in the strict sense retained its connection with the Cachapoal lineage. Historical names thus helped organize the new picture of relationships [Walter et al. 2024].

Species descriptions use combinations of characters. E. robusta has broad, deeply notched ribs and large areoles. E. andicola is notable for slender, two-coloured spines. E. grandiflora has large red flowers, while E. orientalis has dark-purple outer floral parts. Seed comparisons examine tubercles and ridges on the coat. Each element gains meaning when considered with the other characters and the plant’s provenance. Some branches still require better resolution, and the authors point to genomic data as a way forward [Walter et al. 2024].

Geography and routes of pollen transfer

The revised boundaries form a recognizable geographical mosaic. E. mutabilis occupies a narrow coastal area between Los Vilos and Los Molles. E. choapensis occurs inland, while E. andicola and E. grandiflora represent mountain environments. Valleys, ridges and separated patches of suitable habitat can limit contact among populations. The researchers also consider historical climate change. Reconstructing its role requires combining present distributions with a model of lineage history, because current distances between localities describe only one moment in time [Villalobos-Barrantes et al. 2022] [Walter et al. 2024].

Pollen can move by routes different from those of seeds, and its transport depends on animals visiting flowers. E. coimasensis at Las Coimas provides a useful example. In an experiment, hand cross-pollination produced substantially more seeds than self-pollination. Isolated flowers left without hand pollination produced no seeds. Researchers then compared plants beneath Flourensia thurifera shrubs and in the open, using cages excluding hummingbirds and uncaged controls in both environments [Cádiz-Véliz et al. 2021].

Exposed plants accessible to birds had the greatest reproductive output. Beneath shrubs, larger stems accompanied lower reproductive success. The result supports an explanation in which dense cover restricts access by the hummingbird Patagona gigas. The authors also consider possible contributions from nectar quality and conditions for fruit maturation. This experiment shows how fine-scale habitat structure can alter the circumstances of gene transmission. The appearance of a large, well-developed plant provides only part of the information about its contribution to the next generation [Cádiz-Véliz et al. 2021].

A 2025 study combined genomics with pollination ecology in co-occurring cacti. E. chilensis resembled bee-pollinated E. mutabilis, although it was more closely related to hummingbird-visited E. litoralis. The findings support a role for pollinators in reproductive isolation. Proposed floral mimicry, similarity of signals directed at pollinators, remains a hypothesis requiring further experiments [Guerrero et al. 2025].

Name history changes the scope of a result

Research on flower longevity illustrates why locality information must be retained. In a 2022 paper, the montane population at Lomas del Viento, at 2200 m elevation, was identified as E. curvispina. Flowers opened on several consecutive days and closed at night. With stigmas protected from external pollen, the complete opening-and-closing cycle averaged 2.77 days. The cumulative period of full expansion was about 10 hours. Hand cross-pollination left flower longevity similar to that of the controls [Cuartas-Domínguez et al. 2022].

Flowers contained more than 5 g of water on average, with over 2 g in the petal-like tepals. Researchers therefore documented a substantial water stock maintained in a short-lived organ. Fresh and dry mass measurements quantify that stock, whereas estimating the daily cost requires measuring water loss over time. Two bees of the genus Trichothurgus accounted for most visits. In the study season, natural pollination led to fruit formation in 62.2% of observed flowers, compared with 97.6% following intensive hand pollination [Cuartas-Domínguez et al. 2022].

The results retain their value after taxonomic change when accompanied by study locality and the historical scope of the name. The revision assigned Farellones material to E. grandiflora, but transferring detailed findings from another locality requires identification to be confirmed. The Lomas del Viento population therefore remains associated with E. curvispina in its former broad sense. This treatment keeps the experiment useful while preserving precision when comparing currently recognized species [Walter et al. 2024] [Cuartas-Domínguez et al. 2022].

From species boundaries to population conservation

Another dimension of differentiation concerns the beginning of life. An experiment on E. subgibbosa and E. villosa compared seeds from two contrasting habitats. Temperature was combined with water availability adjusted using mannitol. E. subgibbosa showed a higher overall probability of germination, while strongly reduced water potential restricted both species. Sampling represented one population per species, with different seed storage periods. This study adds the germination niche to the picture of diversity, a subject developed further in the article on seed and seedling biology [Guerrero et al. 2016].

After recognizing nine species in the former complex, the revision’s authors proposed threat categories for them. They identified E. aconcaguensis, E. mutabilis and E. orientalis as Critically Endangered, CR, with extents of occurrence below 100 km². These are assessment proposals from the 2024 publication. Dividing a broad name reveals the scale of threats to each lineage, including development, mining, agriculture and fire. In this framework, protecting one large population has a different scope from preserving several distinct evolutionary histories [Walter et al. 2024].

DNA, structural and ecological evidence complement one another at different levels. Markers help recognize lineages and population structure, plant characters allow names to be applied in the field, and pollination and germination experiments reveal processes sustaining successive generations. Accurate records of provenance, observation dates and taxonomic treatment allow this information to be reassessed as research progresses. Together they build an increasingly precise picture of diversity that was long contained within a few broadly applied names [Guerrero et al. 2019] [Villalobos-Barrantes et al. 2022] [Walter et al. 2024] [Guerrero et al. 2016] [Cádiz-Véliz et al. 2021].

Related plant cards