Abstract
<p>Morphological phylogenetic analyses of Macronarian sauropods have expanded substantially over the past two decades, yet relationships among several major clades remain unstable. Although missing data and homoplasy are widely recognised as sources of phylogenetic uncertainty, their relative and combined effects on topological stability have rarely been assessed across successive published datasets. Here, eight influential Macronarian morphological matrices (2002–2026) were reanalysed under maximum parsimony to evaluate the relationships among missing-data architecture, homoplasy, nodal support, and tree stability, as a historical trajectory study. Missing data were quantified using both unknown ('?') entries alone and combined unknown plus gap/inapplicable ('-') entries, while homoplasy was assessed using the Consistency Index (CI) and Retention Index (RI). Topological stability was evaluated using bootstrap support, Bremer support, strict-consensus resolution, Robinson–Foulds (RF) distances, multidimensional tree-space analyses, rogue-taxon detection, and targeted sensitivity analyses. Higher homoplasy was consistently associated with reduced bootstrap support, supporting a negative relationship between character conflict and phylogenetic signal. However, removing highly homoplastic characters improved CI and RI without consistently increasing tree stability, indicating that homoplastic characters may still retain informative signal. In contrast, pruning the most incomplete 10% of taxa produced substantial improvements in consensus resolution, bootstrap support, and RF dispersion, demonstrating that concentrated missing data exert a disproportionate influence on instability. Matrix-specific behaviour further showed that coding gap and inapplicable states as missing information can substantially alter perceived completeness and explain instability not evident from unknown data alone. These findings demonstrate that phylogenetic instability in Macronaria is governed not simply by the quantity of missing data or homoplasy, but by their distribution within matrix architecture and their interaction with phylogenetically influential taxa. Incorporating matrix architecture into future dataset design and taxon sampling will improve the robustness of morphological phylogenetic inference in sauropods and other fossil clades.Keywords: Macronaria; Sauropoda; phylogenetics; missing data; homoplasy; topological instability; Robinson–Foulds distance; tree space; rogue taxa; morphological matrices; parsimony.</p>