# Source paper archive - Source title: Geometric Morphometrics of Pistillate Flowers in Astrocaryum sect. Huicungo (Arecaceae): Taxonomic Implications for the Peruvian Species - Author(s): Valeria Meza, Gabriel Moreno, Marcelo Daza, Betty Millan - Source landing URL: https://www.biorxiv.org/content/10.64898/2026.11.04.764277v1 - PDF URL: https://www.biorxiv.org/content/10.64898/2026.11.04.764277v1.full.pdf - PDF SHA-256: ea3cd65978bb64905b99382d2494c052b412b0e326be0b4c580082a4b398ef49 - Converter: pymupdf4llm 0.0.27 - Limits: Automated PDF text extraction; reading order, equations, tables and formatting may be incomplete or incorrect. Images are omitted, and no OCR was performed. Markup and links are neutralized for safe downloading. Consult the original PDF for the authoritative version. --- [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764277; this version posted November 6, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.04.764277) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 1 _Article_ 2 # **Geometric Morphometrics of Pistillate Flowers in Astrocaryum** 3 sect. Huicungo (Arecaceae): Taxonomic Implications for the 4 # **Peruvian Species** 5 **Valeria Meza** **[1]** **, Gabriel Moreno** **[1]** **, Marcelo Daza** **[1]** **and Betty Millán** **[1,2]** ***** 6 1 Museo de Historia Natural, Universidad Nacional Mayor de San Marcos, Av. Arenales 1256 Jesús María, 7 [Lima, Perú; gabriel.moreno1@unmsm.edu.pe (G.M.); valeria.meza@unmsm.edu.pe (V.M.);](mailto:gabriel.moreno1@unmsm.edu.pe) 8 [marcelodaza@unmsm.edu.pe (M.D.); bmillans@unmsm.edu.pe (B.M.).](mailto:marcelodaza@unmsm.edu.pe) 9 2 Facultad de Ciencias Biológicas, Universidad Nacional Mayor de San Marcos, Ciudad Universitaria, 10 Av. Venezuela Cdra 34 s/n. Apartado 110058, Lima 11, Perú. 11 ***** Correspondence: bmillans@unmsm.edu.pe 12 **Abstract** 13 The taxonomy of the Amazonian palm clade _Astrocaryum_ sect. _Huicungo_ remains 14 controversial due to deep discordances between traditional classifications based on 15 qualitative floral traits and recent phylogenomic evidence. To address this, landmark16 based two-dimensional geometric morphometrics (4 landmarks, 16 semilandmarks) was 17 applied to evaluate the symmetric component of shape variation in calyx and corolla 18 structures across 549 pistillate flowers from the 11 Peruvian species. Although 19 multivariate analyses (PCA, CVA) revealed continuous morphological variation and 20 extensive phenotypic overlap, perianth geometry retained a significant diagnostic signal. 21 Cross-validated classification rates demonstrated that the corolla exhibited higher 22 discriminatory power than the calyx, while the combined perianth analysis (calyx + 23 corolla) maximized taxonomic assignment accuracy. Furthermore, Mahalanobis distance24 based clustering did not replicate traditional subsections: _A. chonta_ and _A. ulei_ exhibited 25 divergent architectures, whereas _A. carnosum_, _A. faranae_, and _A. huicungo_ shared profound 26 geometric similarity. This study demonstrates that quantitative perianth morphometrics 27 is a valuable tool for species delimitation in _Astrocaryum_, supporting the re-evaluation of 28 traditional taxonomic boundaries through an integrative approach. 29 **Keywords:** Arecaceae; palms; geometric morphometrics; floral morphology; 30 classification; integrative taxonomy. 31 32 33 34 35 36 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764277; this version posted November 6, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.04.764277) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 37 **1. Introduction** 38 The Arecaceae family constitutes an indispensable structural and functional component 39 of tropical ecosystems across the Amazon basin and its peripheral regions, where it 40 dominates various forest strata and provides vital resources for both wildlife and local 41 communities **[1,2]** . Within this family, the genus _Astrocaryum_ G. Mey. stands out for its 42 high diversity, harboring approximately 40 species distributed primarily across South 43 America. Section _Huicungo_ (subgenus _Monogynanthus_ ) represents the most diverse lineage 44 within the genus, comprising 15 species, 11 of which have their center of diversification 45 in Peru **[3]** . These palms exhibit a remarkable variety of growth habits, ranging from 46 acaulescent forms to large canopy-reaching arborescent palms, and occupy contrasting 47 ecological niches, from seasonally flooded plains to terra firme forests and Andean 48 foothills **[4,5]** . 49 Despite their ecological importance and hyperdominance, taxonomic delimitation within 50 section _Huicungo_ has been the subject of extensive and ongoing debate. The current 51 infrageneric classification is primarily based on the qualitative evaluation of pistillate 52 flower morphology, utilizing characters such as the length ratio between the calyx and 53 corolla, the presence or absence of indumentum (prickles and bristles), and the 54 architecture of the staminodial ring **[4]** . Based on these attributes, Kahn **[3]** subdivided the 55 section into three subsections: _Huicungo_, _Sachacungo_, and _Murumuru_ . However, current 56 research has revealed strongly divergent hypotheses regarding the evolution of this 57 group. On one hand, phylogenetic and phylogeographic studies have demonstrated a 58 spatial structuring into two major geologically divided clades **[6,7]** . On the other hand, 59 recent phylogenomic analyses based on plastid genomes have refuted the monophyly of 60 the traditional subsections, suggesting that classifications based on qualitative floral 61 morphology might be artificial and that taxonomic boundaries are highly diffuse **[8]** . 62 Given the discordance between phenotype and genotype, combined with the proven 63 inadequacy of classical linear morphometrics to resolve species complexes with high 64 intraspecific variability, geometric morphometrics has emerged as a crucial 65 complementary quantitative approach **[9,10]** . This methodology, based on configurations 66 of Cartesian coordinates (landmarks and semilandmarks), enables the isolation of pure 67 geometric information by mathematically removing the effects of size, position, and 68 rotation **[11]** . In botanical systematics, geometric morphometrics has proven highly 69 effective for delimiting cryptic taxa from herbarium specimens, assessing evolutionary 70 allometry, and quantifying the integration and shape of floral structures, thereby 71 overcoming the inherent subjectivity of historical qualitative descriptions **[12,13]** . 72 Consequently, the main objective of this study is to quantitatively evaluate the 73 morphological variability of the calyx and corolla shape of pistillate flowers in the 74 Peruvian species of _Astrocaryum_ sect. _Huicungo_ using two-dimensional geometric 75 morphometrics. This research seeks to determine whether perianth geometry retains a 76 significant taxonomic signal that improves species delimitation, thus testing the viability 77 of the classical classification system. The principal conclusions of this work reveal that, 78 despite the presence of extensive phenotypic overlap and high floral developmental 79 instability (fluctuating asymmetry), the combined geometry of the calyx and corolla 80 retains a significant diagnostic signal. Furthermore, the recovered phenetic affinities do 81 not replicate traditional subsections, providing quantitative evidence that aligns with 82 recent phylogenomic and biogeographical hypotheses for the taxon. 83 84 **2. Results** [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764277; this version posted November 6, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.04.764277) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 85 **2.1. Shape Variation of the Calyx and Corolla Among Species** 86 The Procrustes analysis of variance to evaluate bilateral symmetry (Table 1) revealed that the 87 highest proportion of morphological variation in the perianth corresponds to the symmetric 88 component (SC), which accounted for 80.07% of the total variance in the calyx and 68.08% in the 89 corolla. The predominance and high statistical significance of this component (𝑝< 0.001) 90 confirm it as the primary source of geometric variation, thereby justifying the exclusive use of 91 the symmetric component for all subsequent analyses. 92 Conversely, fluctuating asymmetry (FA) constituted the second most important factor of 93 variation, capturing 19.39% of the variance in the calyx and 30.53% in the corolla. This proportion 94 evidences a considerable degree of intra-individual developmental instability in response to 95 random perturbations. Finally, although directional asymmetry (DA) accounted for a marginal 96 fraction of the total variance (0.53% for the calyx and 1.38% for the corolla), its effect was 97 statistically significant (𝑝< 0.001), indicating the existence of a subtle yet consistent systematic 98 differentiation between the sides of both structures at the population level. 99 100 **Table 1.** Results of the bilateral symmetry analysis for the calyx and corolla outline. 101 102 103 104 105 **Calyx** 106 **Effect** **Df** **SS** **MS** **%Var** **F** _**p**_ 107 **SC** 548 28.152 0.051373 80.073 4.1296 0.001 108 **DA** 1 0.189 0.18867 0.537 15.1663 0.001 109 110 **FA** 548 6.817 0.01244 19.39 - 111 **Corolla** 112 **Effect** **Df** **SS** **MS** **%Var** **F** _**p**_ 113 **SC** 548 20.1921 0.03685 68.077 2.2295 0.001 114 115 The **DA** 1 0.4117 0.41175 1.388 24.9133 0.001 **Calyx** **Effect** **Df** **SS** **MS** **%Var** **F** _**p**_ **SC** 548 28.152 0.051373 80.073 4.1296 0.001 **DA** 1 0.189 0.18867 0.537 15.1663 0.001 **FA** 548 6.817 0.01244 19.39 - **Corolla** **Effect** **Df** **SS** **MS** **%Var** **F** _**p**_ **SC** 548 20.1921 0.03685 68.077 2.2295 0.001 **DA** 1 0.4117 0.41175 1.388 24.9133 0.001 **FA** 548 9.0569 0.01653 30.535 - 116 evaluated effects correspond to the symmetric component of variation (SC), directional asymmetry (DA), and 117 fluctuating asymmetry (FA). Df, degrees of freedom; SS, sum of squares; MS, mean square; %Var, percentage of 118 variance explained; F, F-value; _p_, p-value. 119 120 To further quantify the relative contribution of taxonomic, intra-specific, and architectural factors to 121 floral shape variation, a hierarchical Procrustes ANOVA was performed (Table 2). The analysis 122 revealed that species identity constituted the primary source of geometric variation, explaining 46.90% 123 and 44.00% of the variance in the calyx and corolla, respectively ( _p_ = 0.001). This confirms that the 124 perianth retained a robust diagnostic taxonomic signal. Notably, intra-specific variability (individuals 125 nested within species) also accounted for a substantial proportion of the variance (31.00% in the calyx 126 and 27.30% in the corolla), quantitatively corroborating the high phenotypic plasticity and 127 developmental instability of these structures. Conversely, the architectural effect—defined by the 128 position of the flower on the rachilla (apical, middle, or basal)—exerted a statistically significant but 129 biologically marginal influence, explaining only 0.60% and 2.80% of the shape variance in the calyx 130 and corolla, respectively. 131 132 133 134 135 136 **Table 2.** Hierarchical Procrustes ANOVA evaluating the effects of species, floral position, and individual 137 variation on the symmetric shape component of the calyx and corolla. [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764277; this version posted November 6, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.04.764277) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 138 **ANOVA - Calyx** **Effect** **Df** **SS** **MS** **%Var** **F** _**p**_ Species 10 6.372 0.637 46.9 7.56 0.001 Individual (Species) 50 4.216 0.084 31 14.64 0.001 Position 2 0.085 0.043 0.6 7.41 0.001 Species * Position 20 0.225 0.011 1.6 1.95 0.002 Residuals 466 2.684 0.006 19.8 - **ANOVA - Corolla** Effect Df SS MS %Var F _**p**_ Species 10 4.453 0.445 44 8.05 0.001 Individual (Species) 50 2.765 0.055 27.3 10.57 0.001 Position 2 0.288 0.144 2.8 27.56 0.001 Species * Position 20 0.178 0.009 1.8 1.7 0.004 Residuals 466 2.439 0.005 24.1 - 139 140 Note: Df, degrees of freedom; SS, sum of squares; MS, mean square; %Var, percentage of variance explained; F, 141 F-value; _p_, p-value. Bold values indicate statistical significance ( _p_ < 0.01). 142 143 The Principal Component Analysis (PCA) of the calyx revealed that the first two 144 components (PC1 and PC2) accounted for 94.41% of the total variation (Figure 1a). Despite 145 a pronounced general overlap, _Astrocaryum ulei_ and _A. chonta_ were distinguished by a 146 clear trend toward positive values on both axes. In the corolla, PC1 and PC2 explained 147 89.21% of the morphological variance (Figure 1b), showing a high degree of morphospace 148 overlap. The combined analysis (calyx + corolla) (Figure 1c) demonstrated that the first 149 two components captured 85.65% of the geometric variation, displaying an ordination 150 pattern consistent with that of the calyx. 151 152 **Figure 1.** PCA plots utilizing the symmetric component of shape, considering all species of the 153 _Huicungo_ Section. Each point represents an individual pistillate flower, and the shaded confidence 154 ellipses are drawn at a 95% probability level. **(a)** Calyx; **(b)** Corolla; **(c)** Calyx + corolla. 155 156 Upon evaluating the symmetric shape changes (Figure 2), variation along PC1 in the calyx 157 represented a gradient from elongated, narrow shapes (associated with negative values) [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764277; this version posted November 6, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.04.764277) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 158 to broad, short structures (positive values). The variation explained by PC2 was 159 concentrated at the base and distal margin; specifically, positive values denoted a 160 widening at the base accompanied by distal narrowing (Figure 2a). 161 Congruently, the variation patterns in the corolla replicated the trends observed in the 162 calyx. PC1 described the same longitudinal and transverse morphological transition (from 163 elongated/narrow to broad/short). Finally, PC2 modeled the transverse variation at the 164 corolla extremities: the negative pole displayed a conspicuously narrow base with a broad 165 distal margin, whereas the positive pole exhibited the inverse conformation (broad base 166 and narrow distal margin) (Figure 2b). 167 **Figure 2.** Symmetric shape changes corresponding to the extreme values (negative and positive) of 168 PC1 and PC2. (a) Calyx; (b) Corolla. 169 **2.2. Mean Shapes of the Calyx and Corolla** 170 The mean shapes of the calyx and corolla, estimated via Procrustes analysis, demonstrated 171 high overall similarity among the 11 studied species, exhibiting only minor deviations 172 from the consensus shapes. Most calyx conformations represented variants of a tubular 173 shape, whereas the corollas tended to be narrower, ranging from tubular to urceolate or 174 obovoid forms (Figure 3). 175 Visualization of the mean calyx shape (Figure 3a) in _Astrocaryum perangustatum_ stood out 176 for exhibiting a clearly tubular morphology with a slender, cylindrical profile, which 177 coincided with the negative extreme of PC1. Similarly, _A. javarense_ and _A. faranae_ 178 presented markedly narrow, elongated calyces; however, _A. javarense_ did not exhibit a 179 perfectly oblong shape. In contrast, _A. ulei_ and _A. chonta_ were characterized by shorter 180 calyces with a broad distal margin. Although both presented an affinity toward 181 cupuliform shapes, _A. chonta_ exhibited a slightly more elongated conformation than _A._ 182 _ulei_ . _A. carnosum_ stood out due to a slightly widened distal margin. Toward the positive 183 extreme of PC1, _A. gratum_ displayed a characteristically urceolate morphology—notably 184 ventricose at the base and slightly narrowed toward the apex. _A. scopatum_ shared this 185 tendency toward a broad base and a narrow orifice, but with a markedly ovoid 186 conformation. Finally, _A. macrocalyx_ showed a suburceolate tendency, presenting an 187 elongated structure that was more ventricose in the medial region than the tubular taxa. 188 In the corolla (Figure 3b), the aspect ratio and lateral curvature allowed for the 189 differentiation of consistent morphological groups. _Astrocaryum carnosum_ and _A. huicungo_ 190 presented highly similar conformations with minimal deviations. As observed in the 191 calyx, _A. perangustatum_ defined the variation extreme toward slender shapes, 192 characterized by an elongated, tubular outline lacking basal expansion—a trend also 193 exhibited by _A. faranae_ . The species _A. ulei_, _A. gratum_, and _A. scopatum_ were characterized [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764277; this version posted November 6, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.04.764277) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 194 by notably broad bases and narrow orifices. Specifically, _A. scopatum_ retained the ovoid 195 shape observed in its calyx, while _A. gratum_ exhibited a pyriform conformation with an 196 evident basal expansion that gradually narrowed toward the apex. _A. urostachys_ presented 197 a globose to ovoid morphology, comparatively broader and with an abruptly truncated 198 apex. Regarding the distal margin, _A. chonta_ maintained a basal urceolate shape but 199 featured the broadest margin of the group. Conversely, _A. macrocalyx_ presented the 200 narrowest distal margin, standing out for its regular, symmetrical outline and a slight 201 subapical constriction. 202 **Figure 3.** Estimated mean shapes for the floral parts of the _Huicungo_ Section species. From left to 203 right: the mean shape calculated using the symmetric component for each species and its deviation 204 from the consensus; graphic representation according to the descriptions by Kahn and Millán [14] 205 and Kahn and Ferreira [15]; and a photograph of a representative calyx for each species (Meza, V. 206 and Moreno, G.). (a) Calyx; (b) Corolla. 207 **2.3. Species Delimitation and Discriminatory Power** 208 The Canonical Variate Analysis (CVA) corroborated the extensive phenotypic overlap 209 among the studied taxa. When evaluating the structures in isolation, the CV1 and CV2 210 axes accounted for 66.93% of the variance for the calyx and 72.18% for the corolla. 211 Conversely, the combined analysis (calyx + corolla) captured 61.01% of the geometric 212 variation (Figure 4). [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764277; this version posted November 6, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.04.764277) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 213 In the specific analysis of the calyx, this high morphospace superimposition precluded an 214 absolute distinction among taxa. Nonetheless, CV2 partially segregated _Astrocaryum_ 215 _chonta_ and _A. ulei_ from the remaining species. 216 Regarding the corolla, despite the high degree of overlap that hindered the unequivocal 217 recognition of the groups, CV1 contributed to the partial separation of _Astrocaryum_ 218 _perangustatum_ from the other species, while CV2 differentiated _A. chonta_ from _A. ulei_ . 219 Finally, the combined analysis (calyx + corolla) also exhibited a high degree of overlap; 220 consistent with the corolla-only analysis, CV1 contributed to the partial segregation of 221 _Astrocaryum perangustatum_ from the remaining taxa. 222 **Figure 4.** CVA plots utilizing the symmetric component of floral part shape for the species of the 223 _Huicungo_ Section. Each point represents an individual pistillate flower, and the shaded confidence 224 ellipses are drawn at a 95% probability level. (a) Calyx; (b) Corolla; (c) Calyx + corolla. 225 The cross-validated classification test derived from the Canonical Variate Analysis (CVA) yielded a 226 correct classification rate (CCR) of 53.92% for the corolla and 44.26% for the calyx. The combined 227 perianth dataset (calyx + corolla) resulted in a CCR of 64.85%. Based on the pairwise Mahalanobis 228 distances evaluated with a Bonferroni correction (Table 3), the calyx recorded significant 229 morphological differentiation ( _p_ < 0.05) across most species comparisons. However, no significant 230 differences ( _p_ - 0.05) were detected between _A. perangustatum_ and _A. urostachys_, nor among 231 _Astrocaryum carnosum_, _A. faranae_, _A. huicungo_, and _A. javarense_ . 232 Regarding the corolla, 9 species pairs lacked significant morphological differentiation ( _p_ - 0.05), 233 including the pairwise comparisons among _A. carnosum_, _A. faranae_, and _A. huicungo_ . Finally, the 234 combined analysis (calyx + corolla) yielded significant Mahalanobis distances for all interspecific 235 comparisons, with the exception of the pairwise comparisons among _Astrocaryum carnosum_, _A._ 236 _faranae_, and _A. huicungo_, which maintained non-significant distances ( _p_ - 0.05) among themselves. 237 Notably, the combined perianth geometry successfully discriminated _A. javarense_ from this group, 238 evidencing highly significant morphometric distances ( _p_ < 0.01) against _A. faranae_ and _A. huicungo_ 239 240 241 242 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764277; this version posted November 6, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.04.764277) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 243 **Table 3.** Pairwise Mahalanobis distances derived from the Canonical Variate Analysis (CVA) of the 244 symmetric shape component for the calyx, corolla, and combined perianth datasets. **Calyx** 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 _**A. car**_ _**A. cho**_ _**A. far**_ _**A. gra**_ _**A. hui**_ _**A. jav**_ _**A. mac**_ _**A. per**_ _**A. sco**_ _**A. ule**_ _**A. cho**_ 1.87** _**A. far**_ 1.15 2.34** _**A. gra**_ 3.04** 3.24** 3.13** _**A. hui**_ 1.19 2.11** 1.24 3.10** _**A. jav**_ 1.56 2.83** 1.41 3.52** 1.78 _**A. mac**_ 2.99** 3.18** 2.90** 3.11** 2.45** 3.24** _**A. per**_ 1.37* 2.80** 1.48** 2.92** 1.50* 1.68* 3.08** _**A. sco**_ 2.95** 2.88** 2.89** 1.48** 2.73** 3.29** 2.57** 2.82** _**A. ule**_ 3.42** 1.83* 3.84** 4.07** 3.70** 4.29** 4.58** 4.30** 3.84** _**A. uro**_ 2.39** 3.12** 2.26** 2.64** 2.35** 2.36** 3.05** 1.89 2.64** 4.24** **Corolla** _**A. car**_ _**A. cho**_ _**A. far**_ _**A. gra**_ _**A. hui**_ _**A. jav**_ _**A. mac**_ _**A. per**_ _**A. sco**_ _**A. ule**_ _**A. cho**_ 1.63* _**A. far**_ 1.31 2.53** _**A. gra**_ 2.75** 3.28** 2.91** _**A. hui**_ 1.37 2.11** 1.27 2.19** _**A. jav**_ 1.61 1.72 2.25** 3.36** 2.05* _**A. mac**_ 2.63** 3.35** 2.33** 2.16** 2.06** 3.78** _**A. per**_ 3.76** 4.10** 4.25** 4.99** 4.30** 4.40** 4.77** _**A. sco**_ 2.74** 3.17** 3.12** 1.38 2.40** 3.44** 2.12** 4.71** _**A. ule**_ 2.94** 2.62** 3.43** 2.01** 2.62** 3.14** 3.05** 4.63** 2.08** _**A. uro**_ 2.25 2.56** 2.62** 2.83** 2.26 2.06 3.25** 4.61** 2.74** 3.11** **Calyx + Corolla** _**A. car**_ _**A. cho**_ _**A. far**_ _**A. gra**_ _**A. hui**_ _**A. jav**_ _**A. mac**_ _**A. per**_ _**A. sco**_ _**A. ule**_ _**A. cho**_ 2.66** _**A. far**_ 1.71 3.45** _**A. gra**_ 3.68** 4.72** 3.73** _**A. hui**_ 2.05 3.68** 1.84 3.47** _**A. jav**_ 2.32 3.40** 2.68** 4.35** 2.74** _**A. mac**_ 3.80** 4.95** 3.45** 3.60** 2.99** 4.74** _**A. per**_ 4.20** 5.30** 4.63** 5.44** 4.56** 5.03** 5.47** _**A. sco**_ 3.57** 4.37** 3.58** 2.23** 3.21** 4.20** 3.17** 5.22** _**A. ule**_ 4.02** 3.22** 4.66** 4.67** 4.54** 4.83** 5.61** 6.14** 4.72** _**A. uro**_ 3.38** 4.05** 3.44** 3.42** 3.37** 3.15** 4.37** 5.12** 3.41** 4.98** 280 _A, car: Astrocaryum carnosum; A. cho: A. chonta; A. far: A. faranae; A. gra: A. gratum; A. hui: A. huicungo;_ 281 _A. jav: A. javarense; A. mac: A. macrocalyx; A. per: A. perangustatum; A. sco: A. scopatum; A. ule: A. ulei;_ 282 _A. uro: A. urostachys._ Asterisks indicate statistical significance after applying a strict Bonferroni 283 correction for 55 pairwise comparisons _(* p_ < 0.05; ** _p_ < 0.01). 284 The cluster analysis based on Mahalanobis distances for the calyx (Figure 5a) 285 revealed that _Astrocaryum ulei_ constitutes the most morphologically divergent taxon, 286 positioning itself isolated from the remaining species. Within the main cluster, two 287 primary groups were distinguished: the first comprised _A. scopatum_, _A. gratum_, and 288 _A. macrocalyx_, united by geometric proximity; the second was characterized by high [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764277; this version posted November 6, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.04.764277) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 289 similarity among _A. carnosum_, _A. huicungo_, and _A. faranae_, forming a subgroup to 290 which _A. perangustatum_, _A. javarense_, _A. urostachys_, and _A. chonta_ adhered successively. 291 Regarding the corolla conformation (Figure 5b), the similarity pattern exhibited 292 notable variations. In this case, _Astrocaryum perangustatum_ was recovered as the most 293 divergent species. The group formed by _A. scopatum_ and _A. gratum_ maintained its 294 proximity, but on this occasion integrated _A. ulei_ and _A. macrocalyx_ . Conversely, the 295 close morphological affinity among _A. carnosum_, _A. faranae_, and _A. huicungo_ remained 296 consistent, forming a cluster associated with _A. urostachys_, while _A. javarense_ and _A._ 297 _chonta_ formed a subgroup close to the latter block. 298 Finally, the dendrogram calculated from the concatenated dataset (calyx + corolla) 299 synthesized the geometric trends of both structures (Figure 5c). _Astrocaryum ulei_ and 300 _A. perangustatum_ consolidated as the most divergent species, branching externally to 301 the rest of the group. The cluster integrated by _A. scopatum_, _A. gratum_, and _A._ 302 _macrocalyx_ was recovered, consistent with the observation for the isolated calyx. 303 Likewise, the high phenotypic affinity among _A. carnosum_, _A. huicungo_, and _A. faranae_ 304 remained invariable, constituting the core of a broader block that sequentially 305 grouped _A. javarense_, _A. urostachys_, and _A. chonta_ . 306 307 **Figure 5.** Similarity dendrogram obtained from Mahalanobis distances. (a) Calyx; (b) Corolla; (c) 308 calyx + corolla. 309 **3. Discussion** 310 **3.1. Symmetry and Developmental Instability in Floral Morphology** 311 A fundamental biological finding derived from the Procrustes superimposition was the 312 quantification of object bilateral symmetry. Although the symmetric component (SC) 313 accounted for the largest proportion of the geometric variance, a high percentage of 314 fluctuating asymmetry (FA) was detected, capturing 19.39% of the variation in the calyx 315 and 30.53% in the corolla. In geometric morphometrics, fluctuating asymmetry represents 316 small, random deviations from perfect symmetry and is considered a direct indicator of 317 instability or developmental noise during ontogeny **[16]** . This level of morphological 318 instability is not unusual in dense inflorescences and has been corroborated in other 319 complex botanical families, such as Ranunculaceae **[17]**, as well as in the foliar variation 320 of the genus _Alnus_, where fluctuating asymmetry reflects high phenotypic plasticity in 321 response to environmental stressors **[18]** . 322 Because fluctuating asymmetry is not genetically determined, its high prevalence 323 indicates a strong influence of microenvironmental factors. Traditional models often [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764277; this version posted November 6, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.04.764277) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 324 attribute this variation to the position of the flower on the rachilla and local resource 325 allocation **[19,20]** . However, our hierarchical ANOVA demonstrated that flower position 326 had a marginal impact, explaining only 0.60% to 2.80% of the shape variance. Instead, this 327 developmental instability is likely driven by physical crowding as the flowers grow 328 tightly packed on the rachilla, forcing them into constrained developmental modules **[21]** . 329 In addition to these ontogenetic pressures, the perianth faces intense ecological 330 interactions. The chewing activity of pollinating beetles, which is extensively documented 331 in the genus **[22]**, imposes severe mechanical stress that could exacerbate the observed 332 asymmetry and plasticity in mature floral stages. Collectively, the convergence of these 333 intrinsic constraints and extrinsic pressures confirms that the final floral shape in this 334 group is a highly plastic trait, thereby contributing to the extensive interspecific overlap 335 documented in the multivariate analyses (PCA and CVA). 336 **3.2. Geometric Variation and Phenotypic Overlap** 337 Geometric morphometric analysis revealed a continuous morphological variation in the 338 perianth of _Astrocaryum_ sect. _Huicungo_ pistillate flowers, characterized by an extensive 339 phenotypic overlap within the morphospace. This high intra- and interspecific variability 340 challenges the discrete categorization of floral shapes traditionally employed in the 341 taxonomy of the group **[4]** . The use of broad qualitative terminology (e.g., tubular, 342 urceolate, dolioliform) may oversimplify floral geometry and underestimate the true 343 variation present in complex taxa **[23]** . In this context, geometric morphometrics has 344 proven to be a superior tool for isolating biological shape and detecting subtle phenotypic 345 variations that linear measurements fail to capture **[9,24,25]** . 346 Our mean shape estimations demonstrate that, while identifiable extreme phenotypes 347 exist—such as the distinctive urceolate calyx of _A. gratum_ or the tubular conformation of 348 _A. perangustatum_ —most species exhibit redundant floral conformations. This aligns with 349 biometric observations of extensive inter- and intrapopulation variation previously 350 reported for the genus **[26]** . 351 **3.3. Taxonomic Delimitation and Phylogenetic Discordance** 352 Despite the continuous variation and visual overlap observed in the Principal Component 353 Analyses (PCA), statistical tests confirmed that perianth geometry retains a significant 354 diagnostic signal. Specifically, the cross-validated classification rates demonstrated that 355 combining the calyx and corolla datasets substantially enhances taxonomic 356 discrimination, achieving the highest correct classification rate (64.85%). This 357 improvement achieved by concatenating multiple morphological structures aligns with 358 documentation in other complex botanical groups, where multi-structure integration 359 more faithfully reflects evolutionary lineages and maximizes phenotypic resolution 360 **[27,28]** . 361 However, the interpretation of these statistical differences requires analytical caution. On 362 one hand, utilizing individual flowers as sampling units can introduce pseudoreplication 363 effects that may artificially inflate statistical significance **[29]** . On the other hand, the high 364 dimensionality inherent to geometric morphometric coordinate datasets can generate 365 apparent mathematical distances even among highly similar morphological distributions 366 **[30,31]** . Therefore, while the pairwise Mahalanobis distances support the separation of 367 most species, these metrics must be understood as mathematical indicators of shape 368 divergence rather than absolute biological barriers. [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764277; this version posted November 6, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.04.764277) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 369 Beyond these analytical considerations, phenetic similarity relationships derived from the 370 UPGMA clustering revealed a clear incongruence with the taxonomic subsections 371 proposed by Kahn **[3]** . A representative case is the high phenetic proximity observed in 372 the UPGMA clustering among _Astrocaryum carnosum_, _A. faranae_, and _A. huicungo_ . 373 However, the interpretation of these phenotypic affinities must decouple graphic 374 clustering from absolute statistical non-significance; indeed, the concatenated pairwise 375 Mahalanobis distances revealed that the combined perianth successfully discriminates _A._ 376 _carnosum_ from its counterparts, whereas non-significant differences persist strictly within 377 the isolated calyx analysis. This shared geometric pattern in the calyx structure points 378 toward limited morphological differentiation or a recent divergence event rather than 379 evolutionary convergence, a scenario highly congruent with recent plastid genome 380 phylogenomics that failed to recover the monophyly of traditional subsections **[7,8]** . 381 Ultimately, this complex morphological continuum among geographically contiguous 382 western Amazonian taxa **[32]** likely reflects a process of recent speciation where ancestral 383 genetic traits are still shared, driven by Late Neogene geoclimatic transitions **[33]** . 384 **4. Materials and Methods** 385 **4.1. Plant material** 386 The botanical material for this study was obtained from herbarium specimens 387 deposited at the San Marcos Herbarium (USM), representing the 11 species of the genus 388 _Astrocaryum_ sect. _Huicungo_ found in Peru ( _Astrocaryum carnosum_, _A. chonta_, _A. faranae_, _A._ 389 _gratum_, _A. huicungo_, _A. javarense_, _A. macrocalyx_, _A. perangustatum_, _A. scopatum_, _A. ulei_, and 390 _A. urostachys_ ) (Figure 6). 391 To guarantee the architectural representativeness of the inflorescences and avoid 392 morphological alterations associated with maturation, individual selection was governed 393 by two strict criteria: (i) possessing complete inflorescences that allowed for the equitable 394 extraction of flowers from the basal, middle, and apical regions of the rachis; and (ii) 395 verifying that the flowers had been collected during anthesis, given that fertilization 396 induces allometric variations in perianth size. In total, 549 pistillate flowers sampled from 397 61 individuals were measured: _N_ = 54 for _Astrocaryum carnosum_ ; _N_ = 54 for _A. chonta_ ; _N_ = 398 90 for _A. faranae_ ; _N_ = 54 for _A. gratum_ ; _N_ = 36 for _A. huicungo_ ; _N_ = 27 for _A. javarense_ ; _N_ = 54 399 for _A. macrocalyx_ ; _N_ = 81 for _A. perangustatum_ ; _N_ = 54 for _A. scopatum_ ; _N_ = 27 for _A. ulei_ ; and 400 _N_ = 18 for _A. urostachys_ . [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764277; this version posted November 6, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.04.764277) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 401 402 403 **Figure 6.** Geographic locations of the studied specimens of _Astrocaryum_ sect. _Huicungo_ in the 404 Peruvian Amazon. 405 406 **4.2. Image Capture and Landmark Digitization** 407 To expose the perianth contour and minimize visual artifacts, a superficial cleaning 408 of the calyx and corolla was performed to remove prickles and bristles, thereby ensuring 409 a clear visualization of the structure's outline. 410 The floral structures were mounted on a support equipped with integrated lighting 411 to eliminate shadows and were photographed alongside a metric scale. The anatomical 412 orientation of each sample was standardized, strictly adhering to the original arrangement 413 of the pedicel relative to the rachis. 414 Photographic captures were performed using a UNITRON Model 13435 trinocular 415 zoom stereo microscope coupled with CaptaVision+ imaging software. To eliminate 416 measurement bias, all images were recorded by a single operator [34]. To represent 417 perianth geometry, a 2D configuration of 20 points (4 landmarks and 16 semilandmarks) 418 was designed. The curvature was modeled utilizing 4 landmarks (points 1, 10, 11, and 20) 419 and 16 semilandmarks (points 2 to 9 and 12 to 19) (Figure 7) [35]. Image file compilation 420 into TPS format and subsequent manual digitization were executed using tpsUtil v.1.81 421 and tpsDig2 v.2.32 software, respectively [36]. [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764277; this version posted November 6, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.04.764277) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 422 423 424 425 **Figure 7.** Configuration of landmarks (light blue) and semilandmarks (red) used to represent 426 the shape of the calyx and corolla of the pistillate flower. 427 428 **4.3. Data Analysis** 429 Spatial processing and statistical analyses were conducted in R version 4.4.3 [37], 430 implementing the specialized packages _geomorph_ version 4.0.10 [9,38] and _Morpho_ version 431 2.12 [39]. Initially, raw coordinate configurations were subjected to a Generalized 432 Procrustes Analysis (GPA) to mathematically remove the effects of scale, position, and 433 orientation [40,41]. To ensure statistical equivalence across the configuration, 434 semilandmarks were mathematically slid along the contour curves applying the 435 minimum bending energy criterion [11,42]. 436 Because _Astrocaryum_ pistillate flowers exhibit bilateral object symmetry in frontal 437 view, shape variation was analytically partitioned [43]. Mirrored copies of the original 438 coordinates were generated, and paired landmarks were relabeled to perform a joint 439 Procrustes superimposition. This procedure allowed for the isolation of fluctuating and 440 directional asymmetry from the symmetric component [16]. This symmetric component 441 was used exclusively for all subsequent inter- and intraspecific analyses. Morphological 442 trends were explored using a Principal Component Analysis (PCA), visualizing the 443 trajectories of geometric change through thin-plate spline (TPS) deformation grids [40]. 444 Morphological dispersion in the morphospace was evaluated by calculating the mean 445 shapes for each species and generating PCA scatter plots with confidence ellipses. To 446 quantify taxonomic differentiation, a Canonical Variate Analysis (CVA) was performed, 447 and the robustness of the recovered morphological groups was assessed by extracting 448 Mahalanobis distances subjected to an iterative permutation test (10,000 iterations) with a 449 Bonferroni correction for multiple comparisons [13,44]. Furthermore, cross-validated 450 classification rates (CCR) were computed to establish the assignment accuracy for each 451 species. Additionally, the Mahalanobis distance matrices were used to construct phenetic 452 similarity dendrograms using the UPGMA clustering algorithm. The entire statistical 453 workflow was applied independently to the calyx and the corolla. Finally, a concatenated 454 dataset (calyx + corolla) was analyzed to evaluate whether combining both floral 455 structures improved taxonomic discrimination within section _Huicungo_ . 456 457 458 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764277; this version posted November 6, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.04.764277) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 459 **5. Conclusions** 460 Geometric morphometric analysis of the perianth in _Astrocaryum_ sect. _Huicungo_ pistillate 461 flowers demonstrates that floral shape exhibits continuous variation and broad 462 interspecific phenotypic overlap. Despite this high variability, the geometric architecture 463 of the calyx and corolla retains a significant diagnostic signal that enables the statistical 464 discrimination of most studied taxa. 465 Phenetic similarity relationships derived from perianth geometry are incongruent with 466 traditional taxonomic subsections, supporting recent molecular evidence that challenges 467 their monophyly. Although the isolated calyx conformation reveals a high geometric 468 similarity among _Astrocaryum carnosum_, _A. faranae_, and _A. huicungo_, the combined 469 perianth analysis successfully discriminates _A. carnosum_ from the other two species. This 470 intricate morphological continuum suggests limited character differentiation or very 471 recent divergence rather than extensive evolutionary convergence, highlighting that some 472 specific boundaries within the current classification remain diffuse. 473 Finally, these findings underscore that taxonomic delimitation in _Astrocaryum_ should not 474 rely exclusively on qualitative shape descriptors. Instead, it requires the integration of 475 geometric morphometrics, ecological attributes, and genomic data within a 476 comprehensive integrative framework to definitively resolve the systematics of this 477 complex Amazonian group. 478 **Author Contributions:** Conceptualization, B.M.; Methodology, G.M. and V.M.; formal analysis, 479 G.M. and V.M.; investigation, G.M., V.M. and M.D.; data curation, G.M., V.M., M.D. and B.M.; 480 visualization, G.M. and V.M.; writing—original draft preparation, G.M. and V.M.; writing—review 481 and editing, B.M.; supervision, B.M. All authors have read and agreed to the published version of 482 the manuscript. 483 **Funding:** This research was supported by the Universidad Nacional Mayor de San Marcos–R. R. 484 N.° 004305-2024-R /UNMSM and Project number B24100911 – PISNFINV- type Project, year 2024. 485 **Data Availability Statement:** The TPS landmark datasets and the R analysis workflow presented in 486 this study are openly available in the Zenodo repository at [https://doi.org/10.5281/zenodo.21176286.](https://www.google.com/search?q=https://doi.org/10.5281/zenodo.21176286) 487 **Acknowledgments:** The authors express their sincere gratitude to the San Marcos Herbarium 488 (USM) and the Department of Gymnosperms and Monocotyledons at the Museo de Historia 489 Natural (MHN-UNMSM) for providing the facilities, infrastructure, and botanical specimens 490 necessary for this study, as most of the research activities were fully developed within these 491 departments. 492 **Conflicts of Interest:** The authors declare no conflicts of interest. 493 494 495 496 497 498 499 500 501 502 503 504 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764277; this version posted November 6, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.04.764277) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 505 **References** 506 1. Balslev, H.; Grandez, C.; Paniagua Zambrana, N.Y.; Møller, A.L.; Hansen, S.L. Palmas (Arecaceae) útiles en los 507 alrededores de Iquitos, Amazonía Peruana. _Rev. Peru. Biol._ **2008**, _15_, 121–132, doi:10.15381/rpb.v15i3.3343. 508 2. ter Steege, H.; Pitman, N.C.A.; Sabatier, D.; Baraloto, C.; Salomão, R.P.; Guevara, J.E.; Phillips, O.L.; Castilho, C.V.; 509 Magnusson, W.E.; Molino, J.-F.; et al. Hyperdominance in the Amazonian Tree Flora. _Science_ **2013**, _342_, 1243092, 510 doi:10.1126/science.1243092. 511 3. Kahn, F. El Género Astrocaryum (Arecaceae). _Rev. Peru. Biol._ **2008**, _15_, 31–48. 512 4. Kahn, F.; Millan, B. The Huicungo Palms. _Univ. Nac. Mayor San Marcos – Fondo Editor. Inst. Rech. Pour Dév._ **2013**, 513 173. 514 5. Roncal, J.; Kahn, F.; Millan, B.; Couvreur, T.L.P.; Pintaud, J.-C. Cenozoic Colonization and Diversification Patterns 515 of Tropical American Palms: Evidence from Astrocaryum (Arecaceae). _Bot. J. Linn. Soc._ **2013**, _171_, 120–139, 516 doi:10.1111/j.1095-8339.2012.01297.x. 517 6. Jimenez-Vasquez, V.; Millán, B.; Machahua, M.; Kahn, F.; Ramirez, R.; Pintaud, J.-C.; Roncal, J. Dry Season 518 Characteristics in Western Amazonia Underlie the Divergence of Astrocaryum Section Huicungo (Arecaceae) and 519 Evaluation of Potential Anatomical Adaptations. _Bot. J. Linn. Soc._ **2017**, _185_, doi:10.1093/botlinnean/box060. 520 7. Roncal, J.; Couderc, M.; Baby, P.; Kahn, F.; Millán, B.; Meerow, A.W.; Pintaud, J.-C. Palm Diversification in Two 521 Geologically Contrasting Regions of Western Amazonia. _J. Biogeogr._ **2015**, _42_, 1503–1513, doi:10.1111/jbi.12518. 522 8. Rivas-Chamorro, M.; Cadenillas, R.; Ge, X.-J.; Jin, L.; Millán, B.; Roncal, J. Testing Species Relationships and 523 Delimitation in the Amazonian Hyperdominant Astrocaryum Section Huicungo (Arecaceae) Using Chloroplast 524 Data from Genome Skimming. _TAXON_ **2023**, _72_, 501–514, doi:10.1002/tax.12928. 525 9. Adams, D.C.; Rohlf, F.J.; Slice, D.E. A Field Comes of Age: Geometric Morphometrics in the 21st Century. _Hystrix_ 526 _Ital. J. Mammal._ **2013**, _24_, doi:10.4404/hystrix-24.1-6283. 527 10. Zelditch, M.L.; Swiderski, D.L.; Sheets, H.D. Introduction. In _Geometric Morphometrics for Biologists_ ; Elsevier, 2012; 528 pp. 1–20 ISBN 978-0-12-386903-6. 529 11. Bookstein, F.L. Landmark Methods for Forms without Landmarks: Morphometrics of Group Differences in 530 Outline Shape. _Med. Image Anal._ **1997**, _1_, 225–243, doi:10.1016/S1361-8415(97)85012-8. 531 12. Chen, Y.; Jabbour, F.; Novikov, A.; Wang, W.; Gerber, S. A Study of Floral Shape Variation in Delphinieae 532 (Ranunculaceae) Using Geometric Morphometrics on Herbarium Specimens. _Bot. Lett._ **2018**, _165_, 368–376, 533 doi:10.1080/23818107.2018.1427145. 534 13. Gardere, M.L.; Dubuisson, J.-Y.; Muller, S.; Savriama, Y. Geometric Morphometrics of Corolla Shape in 535 Campanula (Campanulaceae) from Cabo Verde Archipelago. _Bot. J. Linn. Soc._ **2019**, _191_, 339–352, 536 doi:10.1093/botlinnean/boz055. 537 14. Kahn, F.; Millán, B. Astrocaryum (Palmae) in Amazonia. A Preliminary Treatment. _Bull. Inst. Fr. DÉtudes Andin._ 538 **1992**, _21_, 459–531, doi:10.3406/bifea.1992.1071. 539 15. Kahn, F.; Ferreira, E.J.L. A New Species of Astrocaryum (Palmae) from Acre, Brazil. _Candollea_ **1995**, _50_, 321–328. 540 16. Klingenberg, C.P.; Barluenga, M.; Meyer, A. Shape Analysis of Symmetric Structures: Quantifying Variation 541 Among Individuals and Asymmetry. _Evolution_ **2002**, _56_, 1909–1920, doi:10.1111/j.0014-3820.2002.tb00117.x. 542 17. Zhao, Z.; Du, G.-Z.; Huang, S.-Q. The Effect of Flower Position on Variation and Covariation in Floral Traits in a 543 Wild Hermaphrodite Plant. _BMC Plant Biol._ **2010**, _10_, 91, doi:10.1186/1471-2229-10-91. 544 18. Marković, M.; Vidaković, V.; Popović, Z. A Geometric Morphometrics Approach to the Study of Natural 545 Variations and Hybrid Detection in Populations of Alnus Incana (L.) Moench and Alnus Rohlenae Vít, Douda and 546 Mandák. _Plants_ **2024**, _13_, 993, doi:10.3390/plants13070993. 547 19. Diggle, P.K. Ontogenetic Contingency and Floral Morphology: The Effects of Architecture and Resource 548 Limitation. _Int. J. Plant Sci._ **1997**, _158_, S99–S107. 549 20. Rudall, P.; Stuessy, T.; Mayer, V.; Hörandl, E. Deep Morphology. Toward a Renaissance of Morphology in Plant 550 Systematics. Regnum Vegetabile Vol. 141. _Kew Bull._ **2004**, _59_, 333, doi:10.2307/4115879. 551 21. Diggle, P.K. Modularity and Intra-Floral Integration in Metameric Organisms: Plants Are More than the Sum of 552 Their Parts. _Philos. Trans. R. Soc. B Biol. Sci._ **2014**, _369_, 20130253, doi:10.1098/rstb.2013.0253. 553 22. Listabarth, C. A Survey of Pollination Strategies in the Bactridinae (Palmae). _Bull. Inst. Fr. DÉtudes Andin._ **1992**, _21_, 554 699–714, doi:10.3406/bifea.1992.1082. 555 23. Gardner, A.G.; Gerald, J.N.F.; Menz, J.; Shepherd, K.A.; Howarth, D.G.; Jabaily, R.S. Characterizing Floral 556 Symmetry in the Core Goodeniaceae with Geometric Morphometrics. _PLOS ONE_ **2016**, _11_, e0154736, 557 doi:10.1371/journal.pone.0154736. [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.04.764277; this version posted November 6, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.04.764277) preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in [perpetuity. It is made available under a CC-BY 4.0 International license.](http://creativecommons.org/licenses/by/4.0/) 558 24. Cope, J.S.; Corney, D.; Clark, J.Y.; Remagnino, P.; Wilkin, P. Plant Species Identification Using Digital 559 Morphometrics: A Review. _Expert Syst. Appl._ **2012**, _39_, 7562–7573, doi:10.1016/j.eswa.2012.01.073. 560 25. Viscosi, V.; Fortini, P.; Slice, D.E.; Loy, A.; Blasi, C. Geometric Morphometric Analyses of Leaf Variation in Four 561 Oak Species of the Subgenus Quercus (Fagaceae). _Plant Biosyst. - Int. J. Deal. Asp. Plant Biol._ **2009**, _143_, 575–587, 562 doi:10.1080/11263500902775277. 563 26. Kahn, F.; Gluchy, D. Variation in Morphology of the Pistillate Flowers of Astrocaryum Urostachys (Palmae) in 564 Amazonian Ecuador. _Nord. J. Bot._ **2002**, _22_, 353–360, doi:10.1111/j.1756-1051.2002.tb01384.x. 565 27. Karbstein, K.; Tomasello, S.; Hodač, L.; Dunkel, F.G.; Daubert, M.; Hörandl, E. Phylogenomics Supported by 566 Geometric Morphometrics Reveals Delimitation of Sexual Species within the Polyploid Apomictic Ranunculus 567 Auricomus Complex (Ranunculaceae). _TAXON_ **2020**, _69_, 1191–1220, doi:10.1002/tax.12365. 568 28. Goswami, A.; Smaers, J.B.; Soligo, C.; Polly, P.D. The Macroevolutionary Consequences of Phenotypic Integration: 569 From Development to Deep Time. _Philos. Trans. R. Soc. B Biol. Sci._ **2014**, _369_, 20130254, doi:10.1098/rstb.2013.0254. 570 29. Hurlbert, S.H. Pseudoreplication and the Design of Ecological Field Experiments. _Ecol. Monogr._ **1984**, _54_, 187–211, 571 doi:10.2307/1942661. 572 30. Mitteroecker, P.; Bookstein, F. Linear Discrimination, Ordination, and the Visualization of Selection Gradients in 573 Modern Morphometrics. _Evol. Biol._ **2011**, _38_, 100–114, doi:10.1007/s11692-011-9109-8. 574 31. Rohlf, F. Why Clusters and Other Patterns Can Seem to Be Found in Analyses of High-Dimensional Data. _Evol._ 575 _Biol._ **2021**, _48_, doi:10.1007/s11692-020-09518-6. 576 32. Kahn, F.; Millán, B.; Pintaud, J.-C.; Machahua, M. Evaluación Detallada de La Distribución de Astrocaryum Sec. 577 Huicungo (Arecaceae) En Perú. _Rev. Peru. Biol._ **2011**, _18_, 279–282. 578 33. Hoorn, C.; Wesselingh, F.; ter Steege, H.; Bermudez, M.; Mora, A.; Sevink, J.; Sanmartin, I.; Sanchez Meseguer, A.; 579 Anderson, C.L.; Figueiredo, J.; et al. Amazonia Through Time: Andean Uplift, Climate Change, Landscape 580 Evolution, and Biodiversity. _Science_ **2010**, _330_, 927–931, doi:10.1126/science.1194585. 581 34. Savriama, Y. A Step-by-Step Guide for Geometric Morphometrics of Floral Symmetry. _Front. Plant Sci._ **2018**, _9_, 582 doi:10.3389/fpls.2018.01433. 583 35. Gunz, P.; Mitteroecker, P. Semilandmarks: A Method for Quantifying Curves and Surfaces. _Hystrix Ital. J. Mammal._ 584 **2013**, _24_, 103–109, doi:10.4404/hystrix-24.1-6292. 585 36. Rohlf, F. The Tps Series of Software. _Hystrix Ital. J. Mammal._ **2015**, _26_, doi:10.4404/hystrix-26.1-11264. 586 37. R Core Team _R: A Language and Environment for Statistical Computing_ ; R Foundation for Statistical Computing: 587 Vienna, Austria, 2025; 588 38. Baken, E.K.; Collyer, M.L.; Kaliontzopoulou, A.; Adams, D.C. Geomorph v4.0 and gmShiny: Enhanced Analytics 589 and a New Graphical Interface for a Comprehensive Morphometric Experience. _Methods Ecol. Evol._ **2021**, _12_, 2355– 590 2363. 591 39. Schlager, S. Morpho and Rvcg – Shape Analysis in R. In _Statistical Shape and Deformation Analysis_ ; Zheng, G., Li, 592 S., Szekely, G., Eds.; Academic Press, 2017; pp. 217–256 ISBN 978-0-12-810493-4. 593 40. Klingenberg, C.P. Visualizations in Geometric Morphometrics: How to Read and How to Make Graphs Showing 594 Shape Changes. _Hystrix Ital. J. Mammal._ **2013**, _24_, doi:10.4404/hystrix-24.1-7691. 595 41. Webster, M.; Sheets, H.D. A Practical Introduction to Landmark-Based Geometric Morphometrics. _Paleontol. Soc._ 596 _Pap._ **2010**, _16_, 163–188, doi:10.1017/S1089332600001868. 597 42. Mitteroecker, P.; Schaefer, K. Thirty Years of Geometric Morphometrics: Achievements, Challenges, and the 598 Ongoing Quest for Biological Meaningfulness. _Am. J. Biol. Anthropol._ **2022**, _178_, 181–210, doi:10.1002/ajpa.24531. 599 43. Mardia, K.V.; Bookstein, F.L.; Moreton, I.J. Statistical Assessment of Bilateral Symmetry of Shapes. _Biometrika_ **2000**, 600 _87_, 285–300. 601 44. Rice, W.R. Analyzing Tables of Statistical Tests. _Evolution_ **1989**, _43_, 223–225, doi:10.2307/2409177. 602