Leaf length-to-width ratio links non-destructive area estimation to functional trait variation in the pachymorph bamboo Neosinocalamus affinis across subtropical China.

The leaf length-to-width (L/W) ratio is a simple morphological descriptor that may integrate information about leaf allometry, hydraulic architecture and resource-use strategy, yet its potential to simultaneously improve non-destructive leaf-area estimation and diagnose leaf economics spectrum (LES) position within a single species is rarely tested. We investigated whether L/W ratio captures coordinated variation in leaf morphology, chemistry and photosynthetic performance in Neosinocalamus affinis (Rendle) Keng f. (N. affinis), a sympodial pachymorph bamboo with notably narrow, lanceolate leaves that is economically and ecologically important in subtropical China. We analysed 162 site-level observations of 14 leaf functional traits - leaf area (LA), leaf length (L), leaf width (W), leaf thickness (LT), specific leaf area (SLA), leaf dry matter content (LDMC), SPAD chlorophyll index, maximum photochemical efficiency (Fv/Fm), and leaf carbon (TC), nitrogen (TN) and phosphorus (TP) concentrations with their stoichiometric ratios - collected from 38 county-level sites in seven provinces of southern China. k-means cluster analysis on L/W identified three data-driven shape classes - broader (L/W ≤ 6.05, n = 65), elliptic (6.05 < L/W ≤ 7.18, n = 58) and slender (L/W > 7.18, n = 39) - and a class-specific multiple linear regression (MLR) model had the best predictive performance (R² = 0.937, RMSE = 1.72 cm², AIC = 196.6), outperforming pooled Montgomery, pooled MLR and class-specific Montgomery models. The Montgomery proportionality coefficient (k) increased monotonically from broader (0.598) through elliptic (0.636) to slender leaves (0.663). One-way ANOVA with Tukey HSD detected significant differences among shape classes for LA, L, W, LT, LDMC, TC, TN and TP (all p < 0.05), but not for SLA, SPAD or Fv/Fm. A strong SLA-LDMC trade-off (r = - 0.850, p < 0.001) and N-P coordination (r = 0.590, p < 0.001) were observed, whereas L/W and LA were only weakly related to chemical traits. Environmental variables explained limited trait variation (multiple-regression R² ≤ 0.15), and variance partitioning showed that 30-64% of trait variance occurred within sites. L/W-based classification provides a parsimonious framework that improves non-destructive leaf-area estimation in N. affinis while capturing part - but not all - of the intraspecific LES. Structural and dry-matter traits align with leaf shape, whereas SLA and photosynthetic performance are largely decoupled from morphology. The combined morphological and functional framework offers a practical tool for bamboo ecosystem assessment and clarifies how the LES manifests within a single morphologically plastic species.