Decoupling Taxon Completeness and Phylogenetic Instability in Macronarian Sauropod Morphological Datasets
Abstract
Understanding the relationship between taxon completeness and phylogenetic stability is critical for morphological systematics, particularly in clades like Macronaria, where fragmentary fossil remains frequently complicate phylogenetic inference. This study evaluates whether topological instability within Macronarian sauropods is a predictable function of observational missingness or a reflection of dataset architecture and underlying character conflict.
Eight published morphological matrices spanning 2002 to 2026 were standardized to analyze baseline phylogenetic behavior, calculate continuous stability metrics—Positional Congruence Reduced (PCR) and Leaf Stability Index (LSI)—and perform formal Iterative Positional Congruence Reduction (IterPCR) rogue identification. Across six estimable datasets, rank-correlation analyses revealed no statistically significant relationship between primary missingness and LSI-derived positional instability. Categorical comparisons showed that highly incomplete terminals and IterPCR-selected rogue taxa overlapped by only 17.1% to 40.6%, demonstrating that gross observational completeness and empirical tree-derived stability capture non-equivalent properties.
Experimental deletion of the top 5% and 10% most incomplete taxa frequently improved strict-consensus resolution and reduced Robinson–Foulds (RF) tree-space dispersion beyond matched random controls. However, matrix-level deletion of independently identified formal rogue taxa yielded highly variable results, occasionally deteriorating consensus resolution and increasing dispersion relative to both high-missingness and random treatments. Anatomical auditing confirmed that extremely fragmentary terminals can remain locally constrained when preserving key diagnostic regions, whereas relatively well-scored taxa can exhibit substantial instability owing to character conflict. Consequently, percentage completeness alone is an inadequate proxy for phylogenetic reliability, and matrix refinement requires evaluating character architecture alongside tree-derived stability.
Related
- MacronariaMatrix: a curated cross-matrix database of morphological phylogenetic characters in macronarian sauropods
- Historical inheritance and expansion of morphological character sampling in macronarian sauropod phylogenetics
- Morphological and Phylogenetic Characterization of Five Unrecorded Endophytic Fungal Species from Woody Plants in Korea
- Emended description of the genus Henriciella: morphological and phylogenetic analysis identifies two subgenera with different reproductive strategies
- Morphological and multilocus phylogenetic analysis of Sirococcus (Diaporthales, Gnomoniaceae) on woody hosts from Fujian, Shandong, and Yunnan, China: Two new species and two new records
BibTeX
@misc{dookhy2026decoupling,
title = {{Decoupling Taxon Completeness and Phylogenetic Instability in Macronarian Sauropod Morphological Datasets}},
author = {Chetan Dookhy},
howpublished = {PaleorXiv},
year = {2026},
month = sep,
doi = {10.31233/osf.io/dbtm3_v2},
url = {https://osf.io/preprints/paleorxiv/dbtm3_v2/},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Decoupling Taxon Completeness and Phylogenetic Instability in Macronarian Sauropod Morphological Datasets
AU - Chetan Dookhy
PY - 2026
DA - 2026/09/22
PB - PaleorXiv
DO - 10.31233/osf.io/dbtm3_v2
UR - https://osf.io/preprints/paleorxiv/dbtm3_v2/
AB - Understanding the relationship between taxon completeness and phylogenetic stability is critical for morphological systematics, particularly in clades like Macronaria, where fragmentary fossil remains frequently complicate phylogenetic inference. This study evaluates whether topological instability within Macronarian sauropods is a predictable function of observational missingness or a reflection of dataset architecture and underlying character conflict. Eight published morphological matrices spanning 2002 to 2026 were standardized to analyze baseline phylogenetic behavior, calculate continuous stability metrics—Positional Congruence Reduced (PCR) and Leaf Stability Index (LSI)—and perform formal Iterative Positional Congruence Reduction (IterPCR) rogue identification. Across six estimable datasets, rank-correlation analyses revealed no statistically significant relationship between primary missingness and LSI-derived positional instability. Categorical comparisons showed that highly incomplete terminals and IterPCR-selected rogue taxa overlapped by only 17.1% to 40.6%, demonstrating that gross observational completeness and empirical tree-derived stability capture non-equivalent properties. Experimental deletion of the top 5% and 10% most incomplete taxa frequently improved strict-consensus resolution and reduced Robinson–Foulds (RF) tree-space dispersion beyond matched random controls. However, matrix-level deletion of independently identified formal rogue taxa yielded highly variable results, occasionally deteriorating consensus resolution and increasing dispersion relative to both high-missingness and random treatments. Anatomical auditing confirmed that extremely fragmentary terminals can remain locally constrained when preserving key diagnostic regions, whereas relatively well-scored taxa can exhibit substantial instability owing to character conflict. Consequently, percentage completeness alone is an inadequate proxy for phylogenetic reliability, and matrix refinement requires evaluating character architecture alongside tree-derived stability.
ER -