This dataset is an alignment of seven plastid markers. These were generated from 156 specimens of Megalastrum, and represent 72 of the 93 species that are currently recognized. The alignment also includes 46 species from 14 other Dryopteridaceae genera. The alignment was generated using MAFFT. Maximium likelihood phylogenies were were generated  using IQ-tree v2.1.3 and MrBayes v.3.2.6. In agreement with previous findings, results indicate that Megalastrum is monophyletic and sister to Rumohra. We recognize six main clades: 1) The Mascarene Clade, with two species endemic to the Mascarenes and having echinate-fenestrate spores; 2) the Pulverulentum Clade, consisting of ten species typically with large, 3–4-pinnate leaves and strongly denticulate scales often darkened apically; 3) the Cristate-spored Clade, harboring 57 species with cristate spores; 4) the Villosum Clade, with a single species endemic to the Greater Antilles (M. villosum, the type of the genus); 5) the Circumaustral Clade, consisting of seven species occurring at high (33°–45°) southern latitudes; and 6) the Southeastern Brazil Clade, containing 16 species mostly or entirely restricted to the region.

Taxon sampling—We included sequences from 156 Megalastrum specimens . These represent 72 of the 93 species (77%) recognized in the genus. As outgroups, we used 46 species from 14 other Dryopteridaceae genera (Appendix 1), including Lastreopsis, Parapolystichum, and Rumohra, which are most closely related to Megalastrum and classified together in the Lastreopsidoideae (Zuo et al. 2025). For the ingroup and outgroup samples sequenced and included in this study, voucher information and GenBank accession numbers are provided in Appendix 1 of the associated manuscript.

For five specimens, the species names used here differ from those provided in our monographs, where they were incorrectly cited. The first is Megalastrum rupicola (Kessler et al. 11621) from Bolivia, which was incorrectly cited by Moran et al. (2014b) as “M. praetermissum”; the second is M. organense (Labiak et al. 4371) from Brazil, which was incorrectly cited by Moran et al. (2009a) as “M. umbrinum”; the last three are M. oppositum from the Mascarenes, which were incorrectly cited by Rouhan & Moran (2011) as “M. canacae” (Duan & Zhang 2014). The correct names are used in Fig. 2 and Appendix 1.

DNA extraction, amplification, and sequencing—DNA was isolated from either herbarium specimens or leaves collected in the field and dried in silica-gel (Appendix 1). To avoid contamination, we collected young leaves free of epiphyllous bryophytes and obvious fungal or insect parasites. For silica-gel dried samples, total genomic DNA was extracted from approximately 1 cm² of leaf tissue using a modified CTAB protocol based on Doyle & Doyle (1987; see Struwe et al. 1998). For herbarium material, the DNEasy Plant Mini kit Qiagen (Valencia, CA) was used following the manufacturer's protocol, but with a proteinase K digestion during the lysis step: 30 μL of proteinase K was added per tube, and tubes were incubated on a tipping plate at 42 °C for 24h.

The polymerase chain reaction (PCR) was used to amplify seven plastid regions: two coding genes (atpA and rbcL) and five non-coding intergenic spacers (psbA-trnH, atpB-rbcL, trnG-trnR, trnL-trnF, and rps4-trnS). PCR reactions were typically carried out in a 25 μL volume containing 1x PCR buffer, 2.5 mM MgCl2, 250 μM of each dNTP, 1 M betaine, 0.4 μM of each primer, 0.75 U Taq polymerase (Taq CORE kit; MP Biomedicals, Illkirch, France), and 0.75 μL of template DNA. Primer sequences and thermal cycling conditions are reported in Table 1.

The resulting PCR products were visualized on a 1 % agarose gel with ethidium bromide and purified using QIAquick PCR purification kit (Qiagen). They were directly sequenced with a Perkin Elmer ABI 377XL automated sequencer using the amplification primers and additional internal primers for some markers (Table 1). Forward and reverse sequences obtained were edited and assembled using Sequencher 4.9 (Gene Codes Corporation, Ann Arbor, MI, USA). The newly obtained consensus sequences were submitted to GenBank, and the alignments to Dryad (Moran et al. 2026); the accession numbers for both repositories are listed in Appendix 1.

Sequence alignment, indel coding, and phylogenetic analyses—We assembled a data matrix of two gene regions (atpA, rbcL) and five intergenic spacers (psbA-trnH, trnG-trnR, trnL-trnF, psbA-trnH, and rps4-trnS). Sequences for each region were aligned using the MAFFT (Katoh et al. 2002) plugin in Geneious 2024.07. Maximum likelihood (ML) phylogenetic analyses were performed using IQ-tree v2.1.3 (Nguyen et al. 2015; Minh et al. 2020). We implemented ModelFinder in IQ-tree (Kalyaanamoorthy et al. 2017) to select the best-fitting model of sequence evolution via Bayesian Information Criterion (BIC). Node support values were generated from 1000 ultrafast bootstrap (UFBoot) replicates (Hoang et al. 2018; Minh et al. 2013). Bayesian inference (BI) of phylogeny was performed using MrBayes v.3.2.6 (Ronquist & Huelsenbeck 2003) and implemented in the CIPRES portal (Miller et al. 2010), with four chains run for 6 million generations, sampling every 1000 generations. The resulting log files were inspected for convergence and adequate sampling using Tracer v.1.7.2 (Rambaut et al. 2018). The first 25% of trees was discarded as burn-in, and a majority-rule consensus tree was generated from the remaining trees. Support values shown are BI posterior probabilities (PP) and ML bootstrap (BS) percentages. The aligned data matrix is available in the Dryad Digital Repository (Moran et al. 2026). In the tree presented (Fig. 2), outgroups were omitted to save space and make the species names easier to read.