Ontogenetic Stage Strongly and Differentially Influences Leaf Economic and Stomatal Traits Along Phyllotactic and Environmental Gradients

Abstract

Phyllotaxy is a key determinant of intraspecific variation in leaf functional traits, with different leaflet positions often representing distinct strategies of resource acquisition and utilization. Yet, the extent to which such phyllotactic differentiation is modulated by ontogenetic stage remains poorly understood. Here, we examined saplings and adult trees of Fraxinus mandshurica, a dominant compound-leaved species in temperate broadleaf forests, by quantifying four leaf economic traits and four stomatal traits across six phyllotactic positions. We further assessed the relative influences of phyllotaxy and environmental factors, including soil total nitrogen, soil water content, and canopy openness, on trait variation at different ontogenetic stages. Our results showed that economic traits varied significantly along phyllotaxy, whereas stomatal traits were relatively conservative. The effects of ontogenetic stage on traits at a given phyllotactic position were trait-specific. Within-group correlations of economic traits and of stomatal traits remained stable across ontogenetic stages and were consistently stronger than between-group correlations. Sapling traits were more strongly affected by soil total nitrogen and soil water content, whereas those in adult trees were primarily shaped by soil water content and canopy openness. Moreover, both trait–trait and trait–environment associations were weaker at the leaflet level than at the compound-leaf level. Our study highlights the critical role of ontogenetic stage in shaping leaf trait responses to phyllotaxy and environmental change, providing new insights into the mechanisms underlying intraspecific trait variation in compound-leaved tree species.

1. Introduction

Variation in leaf functional traits reflects adaptive strategies that enable plants to cope with environmental heterogeneity [1,2,3]. Such variation regulates physiological processes at the individual level and shapes species distribution and functional diversity at the community level [4,5]. While interspecific and community-level patterns of trait variation have been extensively studied, increasing evidence indicates that intraspecific variation constitutes a substantial component of total variation and is critical for understanding plant adaptive strategies and predicting ecosystem functioning [6,7]. The ontogenetic stage, which refers to discrete life-history categories along an individual’s development (e.g., saplings to adult trees), is a key determinant regulating intraspecific variation in leaf traits and their interrelationships [8]. Saplings typically exhibit higher specific leaf area (SLA) and leaf nitrogen content (LN) to support rapid growth and carbon acquisition, whereas adult trees invest more in leaf thickness (LT) and structural toughness, thereby extending leaf lifespan and enhancing tolerance to drought, herbivory, and physical disturbance [8,9,10]. Yet it remains unclear whether, and to what extent, ontogenetic stage modifies the influence of within-individual spatial heterogeneity on patterns of leaf trait variation.

Compound leaves may provide an opportunity to investigate how leaflet development and microenvironmental heterogeneity drive trait variation [11,12]. Along the rachis from apex to base, apical leaflets are generally exposed to stronger light and higher transpirational demand, whereas basal leaflets, being closer to vascular tissues, benefit from more stable supplies of water and nutrients [7,13]. These positional differences form clear within-compound-leaf gradients of light, water, and nutrient availability [14]. Previous studies have shown that compound-leaved species exhibit significant phyllotactic variation in leaf traits, with apical leaflets often having higher SLA and stomatal density (SD) to support carbon assimilation, whereas basal leaflets tend to invest more in LT and nutrient accumulation [15,16]. However, it remains unclear whether phyllotaxis-driven trait differentiation conforms to the broader trade-offs of the leaf economics spectrum.

The influence of environmental factors on leaf traits is, to some extent, mediated by ontogenetic stage [17,18]. Saplings, at an early ontogenetic stage, generally have high growth demands and are characterized by thinner leaves and shallow root systems that limit their capacity to store and buffer water and nutrients, as a result, they are particularly sensitive to fluctuations in soil nutrient availability and water supply [19,20]. For instance, increases in soil nitrogen often promote higher SLA and LN in saplings, accompanied by enhanced stomatal responsiveness to environmental variation, which sustains efficient gas exchange, accelerates photosynthesis, and facilitates leaf area expansion [21,22]. In contrast, adult trees, at a later ontogenetic stage, typically develop thicker leaves with higher tissue density that enhance mechanical support and defense [23,24]. Their deeper root systems provide greater stability and buffering capacity in resource acquisition, thereby reducing sensitivity to short-term fluctuations in nitrogen and water availability [25]. Within vertically stratified canopies, leaf traits in adult trees are more strongly shaped by interacting gradients of light and water heterogeneity [26,27]. Nevertheless, it remains uncertain whether the effects of phyllotaxy and environmental factors on economic and stomatal traits differ across ontogenetic stages.

In this study, we focused on Fraxinus mandshurica, a representative compound-leaved broadleaf species distributed in the zonal climax mixed broadleaved–Korean pine (Pinus koraiensis) forest of Northeast China. We measured four economic traits and four stomatal traits across phyllotactic positions within compound leaves of both saplings and adult trees, and quantified environmental variables, including soil total nitrogen (STN), soil water content (SWC), and canopy openness (CO). Our aim was to assess the effects of phyllotaxy and environmental factors on leaf traits across ontogenetic stages. Specifically, we addressed the following questions: (1) Do leaf economic and stomatal traits vary significantly along phyllotaxy? (2) Does ontogenetic stage (saplings vs. adult trees) regulate trait values and their correlations? (3) Are the effects of phyllotaxy and environmental factors on leaf traits dependent on ontogenetic stage?

2. Materials and Methods

2.1. Study Site

The study area is located in the Muling Taxus cuspidata National Nature Reserve in northeastern China, geographically positioned between 130°00′–130°28′ E and 43°49′–44°06′ N. The reserve is characterized by a zonal climax vegetation of mixed broadleaved–Korean pine forest. Elevation ranges between 500 and 700 m, and soils are classified as dark brown soils. The local climate belongs to the northern temperate continental monsoon type, marked by severe, dry winters, warm and wet summers, and sharp temperature fluctuations during transitional seasons. On average, the site experiences 2.8 °C in annual temperature and 530 mm of rainfall annually, and most of which falls from June to August.

2.2. Sample Collection

In mid-July to August 2024, we randomly selected 20 Fraxinus mandshurica individuals, including 10 saplings and 10 adult trees. Within each ontogenetic stage, sampled individuals were similar in size: saplings had a mean diameter at breast height (DBH) of 4.72 ± 0.21 cm and mean height of 7.06 ± 0.25 m, whereas adult trees had a mean DBH of 21.92 ± 2.43 cm and mean height of 33.83 ± 3.49 m. All individuals grew under comparable slope aspects and gradients. To reduce spatial autocorrelation, individual trees were sampled at intervals of no less than 10 m. A randomly chosen branch from the south-facing upper canopy was collected for each tree by a professional climber [28]. From each sampled branch, 15–20 compound leaves that were mature, fully expanded, and morphologically intact were collected to quantify leaf economic traits and stomatal traits. Collected leaf samples were immediately sealed in plastic bags and delivered to the laboratory. Each compound leaf was dissected by cutting all leaflets from the rachis with scissors and labeling them individually. For each position, the leaflets were evenly divided into three subsamples (each containing approximately 20 leaflets): one for measuring SLA, LDMC, and LT, one for stomatal traits, and one for LN. Measurements of economic traits were completed within 6 h, and leaves intended for stomatal analyses were fixed in FAA solution (a mixture of 70% ethanol, formalin, and glacial acetic acid in a 90:5:5 ratio). Because leaflets beyond the phyllotactic position-6 were often absent or damaged, especially in saplings, and trait variation in both leaf economic and stomatal characteristics became minimal beyond the sixth position [7], only the first six positions were retained for analysis (Figure A1).

For soil collection, surface litter was cleared in a 1 m radius around each tree prior to sampling. Soil samples were then taken from the upper 0–10 cm layer at three equidistant points (120° apart). The three subsamples were thoroughly homogenized, combined into a single composite sample, and transported to the laboratory for nutrient analyses [29].

2.3. Leaf Trait Measurements

For economic trait measurements, three leaflets were randomly selected from the pooled sample at each position for analysis. Fresh mass was measured at a resolution of 0.0001 g. Leaf area was obtained by scanning with a flatbed scanner (BenQ Corp., Taipei, China) and analyzing the images in Photoshop CS 8.01 (Adobe Systems Inc., San Jose, CA, USA) with an accuracy of 0.01 cm². LT was calculated as the mean of three micrometer readings (excluding the midrib; precision 0.01 mm). Samples were dried in an oven at 65 °C for ≥72 h to a constant mass., and dry mass was determined (precision 0.0001 g). Leaf dry matter content (LDMC) was obtained as the proportion of dry to fresh mass, while SLA was defined as the ratio of leaf area to dry mass. For LN, dried and finely ground samples were digested for 40 min in an H₂SO₄–H₂O₂ pre-digestion system, and nitrogen concentration was quantified using a Kjeldahl nitrogen analyzer (Hannon K9840, Jinan Hannon Instrument Co., Ltd., Jinan, China).

For stomatal traits, we employed the nail polish impression technique. Three leaflets were randomly selected from the pooled sample preserved in FAA solution, then removed and air-dried prior to measurement. A thin layer of transparent nail polish, approximately 1 mm² in size, was applied to the abaxial surface. After drying, the film was carefully stripped off with fine tweezers and mounted as a temporary preparation for microscopic observation. Images were taken under a microscope equipped with a 20× objective lens. For each impression, three fields of view were randomly chosen, and stomata were counted. In addition, Image J software (version 1.54f) was used to measure stomatal length (SL) and width (SW) from three randomly selected stomata per field, and mean values were used in subsequent analyses. SD was calculated as the number of stomata divided by the area of the observed photo (mm²) [30]. The stomatal pore index (SPI) was calculated according to the formula: SPI = SL² × SD × 10⁻⁴. The leaf traits information for different ontogenetic stages is provided in

Table 1. Information on leaf economic and stomatal traits across ontogenetic stages.

Ontogenetic Stage	Leaf Traits	Abbreviation	Units	Min	Max	Mean	SE	CV (%)
Sapling	Specific leaf area	SLA	cm2/g	72.685	331.003	201.844	8.281	0.780
	Leaf dry matter content	LDMC	g/g	2.397	3.938	3.168	0.050	0.391
	Leaf nitrogen content	LN	mg/g	30.271	58.533	44.402	0.816	0.483
	Leaf thickness	LT	mm	0.767	2.167	1.467	0.051	0.646
	Stomatal length	SL	μm	11.642	20.586	16.114	0.235	0.434
	Stomatal width	SW	μm	7.390	18.915	13.153	0.266	0.609
	Stomatal density	SD	stoma/mm2	101.423	164.375	132.899	2.072	0.383
	Stomatal pore index	SPI	%	2.181	5.176	3.678	0.102	0.579
Adult tree	Specific leaf area	SLA	cm2/g	61.158	166.230	113.694	2.999	0.632
	Leaf dry matter content	LDMC	g/g	2.019	3.328	2.673	0.042	0.393
	Leaf nitrogen content	LN	mg/g	34.390	51.002	42.696	0.552	0.326
	Leaf thickness	LT	mm	0.733	2.183	1.458	0.048	0.664
	Stomatal length	SL	μm	14.916	21.018	17.967	0.147	0.290
	Stomatal width	SW	μm	10.955	18.670	14.812	0.173	0.413
	Stomatal density	SD	stoma/mm2	87.433	153.883	120.658	2.064	0.432
	Stomatal pore index	SPI	%	3.361	6.749	5.055	0.075	0.502

.

2.4. Environmental Factors Measurements

For CO, three hemispherical photographs were taken for each tree using a Nikon Coolpix 4500 digital camera fitted with a 180° fisheye lens (Nikon Corp., Tokyo, Japan), positioned in both the northern and southern directions. The images were subsequently analyzed with Gap Light Analyzer software (version 2.0) to quantify canopy openness, providing an estimate of local light conditions [31].

SWC was determined gravimetrically as the ratio of oven-dry to fresh mass. STN concentration was measured with a Kjeldahl nitrogen analyzer (Hanon K9840, Jinan Hanon Instruments Co., Ltd., Jinan, China).

2.5. Data Analysis

Statistical computations were carried out in R version 4.2.1. At each phyllotactic position, SLA, LDMC, LT, and stomatal traits were measured three times. However, because the amount of material from a single leaflet was insufficient for accurate chemical analysis, LN was measured only once per position. Therefore, the mean of the three replicate measurements was used for all other traits in subsequent analyses. In addition, since the left and right leaflets at the same position were defined as belonging to the same phyllotactic position, values from the left and right sides were averaged, except for the first position (Figure 1). Differences in traits among phyllotactic positions were evaluated using the Least Significant Difference (LSD) test, which was also applied to assess trait variation between ontogenetic stages. Relationships among leaf traits were examined with Spearman’s rank correlations, after log-transforming all trait data to meet normality assumptions. In the above steps, 60 data points were analyzed for each ontogenetic stage (10 trees × 6 phyllotactic positions). To explore multivariate patterns of leaflet traits across six phyllotactic positions within compound leaves, principal component analysis (PCA) was conducted for morphological and stomatal traits. A general linear model (GLM) was applied to evaluate the effects of phyllotaxy and environmental factors (soil total nitrogen, soil water content, and canopy openness) on eight leaf traits. Analyses were performed in R 4.3.1 (R Core Team, 2023) using the glm () function in the stats package. Phyllotaxy and the three environmental variables were treated as fixed factors, while tree identity was included as a random factor to account for within-tree dependence among leaflets. Prior to analysis, all continuous predictors were standardized (z-scores), and response variables were log-transformed when necessary to meet assumptions of normality and homoscedasticity. To prevent collinearity from biasing model estimates, variance inflation factors (VIFs) were calculated for all predictor pairs; VIF values below 10 indicated that multicollinearity was not a concern. Figures and tables were prepared using R 4.2.1, SigmaPlot 10.0, and Excel 2016.

3. Results

3.1. Variation in Leaf Economic and Stomatal Traits Across Phyllotaxy and Ontogenetic Stages

Variation in leaf traits along phyllotaxy depended on both specific traits and ontogenetic stage (Figure 1). Economic traits exhibited greater variation than stomatal traits: SLA showed no significant differences among phyllotactic positions in either ontogenetic stage (Figure 1A); LDMC decreased and LN increased from apical to basal phyllotactic positions, with similar trends across stages (Figure 1B,C); LT declined with phyllotaxy in adult trees but remained stable in saplings (Figure 1D). Across ontogenetic stages, adults consistently had lower SLA and LDMC (except SLA at phyllotactic position 3), whereas LT and LN showed no significant differences between ontogenetic stages (Figure 1A–D).

Stomatal traits were comparatively conservative (Figure 1E–H). No significant differences in SL, SW, or SPI were observed among phyllotactic positions, while SD decreased from apical to mid positions and increased basally, with minima at intermediate positions in both stages (Figure 1G). Between ontogenetic stages, SL was higher in adult trees than in saplings at all phyllotactic positions except position 6; SW was significantly higher in adults at positions 1, 2, and 5; SD was higher in saplings only at position 1; and SPI was significantly higher in adults at positions 2–4 (Figure 1E–H).

3.2. Correlations Among Leaf Traits Across Ontogenetic Stages

For both ontogenetic stages, at compound-leaf scale, within-group correlations among economic or stomatal traits were largely consistent: SLA showed positive correlations with LDMC and LN, but a negative correlation with LT; SPI showed positive correlations with both SL and SD, whereas SD was negatively related to SL and SW (Figure 2). However, cross-group correlations differed between ontogenetic stages: in saplings, SD exhibited a positive association with SLA and a negative association with LN, whereas these correlations were not significant in adult trees (Figure 2). Overall, correlations among traits at compound-leaf scale within each ontogenetic stage were stronger than those analyzed at leaflet (phyllotactic) scale (Figure 2;

Table A1. Pearson correlation coefficients between leaf traits of sapling and adult at different phyllotactic positions. The light gray areas represent adult trees, the white areas represent sapling trees, and bold values indicate significant correlations at the 0.05 level.

Phyllotaxy		LT	LDMC	SLA	LN	SW	SL	SD	SPI
1	LT		−0.45	−0.36	0.25	−0.03	−0.31	0.24	−0.11
	LDMC	−0.65		0.48	0.03	0.04	0.46	−0.5	−0.1
	SLA	−0.87	0.87		0.54	−0.52	−0.15	0.05	−0.26
	LN	−0.37	0.43	0.31		−0.38	−0.24	0.27	0.01
	SW	−0.16	0.34	0.16	−0.02		0.27	−0.29	−0.02
	SL	−0.09	0.24	0.09	−0.46	0.52		−0.85	0.2
	SD	−0.1	0.26	0.25	0.02	0.32	0.01		0.34
	SPI	−0.05	0.29	0.11	−0.43	0.57	0.93	0.37
2	LT		−0.38	−0.42	−0.09	0.3	0.13	−0.19	−0.14
	LDMC	−0.54		0.45	0.36	0.27	0.26	−0.33	−0.22
	SLA	−0.78	0.79		0.69	0.04	0.17	−0.22	−0.19
	LN	−0.12	0.5	0.35		0.36	−0.07	−0.13	−0.26
	SW	0.06	0.01	−0.35	0.35		0.64	−0.43	−0.02
	SL	−0.54	−0.01	0.11	0.12	0.38		−0.66	−0.01
	SD	−0.19	0.35	0.37	−0.46	−0.65	−0.49		0.75
	SPI	−0.72	0.3	0.44	−0.28	−0.19	0.63	0.37
3	LT		−0.3	−0.43	−0.29	0.31	0	−0.15	−0.24
	LDMC	−0.78		0.82	0.32	0.33	0.42	−0.01	0.45
	SLA	−0.67	0.78		0.45	0.33	0.22	0.06	0.32
	LN	0.02	0.47	0.41		0.48	0.33	−0.3	0.01
	SW	0.03	−0.1	−0.3	−0.07		0.46	−0.46	−0.09
	SL	0.08	−0.32	−0.54	−0.54	0.38		−0.55	0.4
	SD	−0.16	0.29	0.28	−0.05	−0.36	0.35		0.54
	SPI	−0.05	0.03	−0.09	−0.31	−0.03	0.76	0.87
4	LT		−0.49	−0.63	−0.39	−0.12	−0.01	0.24	0.2
	LDMC	−0.82		0.88	0.36	−0.09	−0.32	−0.03	−0.29
	SLA	−0.63	0.82		0.45	−0.02	−0.21	−0.3	−0.49
	LN	−0.25	0.48	0.55		0.11	−0.18	0.17	−0.02
	SW	0.4	−0.32	−0.38	−0.2		0.78	−0.52	0.23
	SL	0.04	−0.31	−0.5	−0.31	0.7		−0.45	0.49
	SD	−0.09	0.27	0.58	−0.04	−0.38	−0.5		0.55
	SPI	0.06	−0.23	−0.21	−0.38	0.57	0.82	0.09
5	LT		−0.42	−0.08	0.02	−0.34	−0.12	−0.59	−0.67
	LDMC	−0.45		0.72	0.18	0.41	0.48	0.17	0.48
	SLA	−0.73	0.3		0.04	0.42	0.36	−0.22	0.01
	LN	−0.24	0.52	0.35		0.38	−0.07	0.23	0.16
	SW	0.27	0.19	−0.39	−0.02		0.46	−0.16	0.18
	SL	0.3	−0.57	−0.42	−0.09	0.3		−0.31	0.42
	SD	0.03	−0.48	0.15	−0.46	−0.54	0.06		0.73
	SPI	0.23	−0.69	−0.27	−0.22	0.07	0.93	0.4
6	LT		−0.61	0.29	−0.57	−0.43	−0.32	0.7	0.31
	LDMC	−0.72		0.39	0.47	0.73	0.64	−0.55	0.08
	SLA	−0.63	0.58		−0.32	0.54	0.52	−0.26	0.24
	LN	−0.25	0.51	0.53		0.34	−0.03	−0.31	−0.33
	SW	0.46	−0.23	−0.26	0.01		0.72	−0.72	0.05
	SL	0.48	−0.23	−0.26	0.05	0.75		−0.38	0.61
	SD	−0.21	0.04	0.39	−0.42	−0.53	−0.53		0.49
	SPI	0.43	−0.31	−0.08	−0.23	0.53	0.85	−0.01

SLA: Specific leaf area; LDMC: Leaf dry matter content; LN: Leaf nitrogen content; LT: Leaf thickness; SL: Stomatal length; SW: Stomatal width; SD: Stomatal density; SPI: Stomatal pore index.

).

Principal component analysis (PCA) indicated that the first two axes explained a similar proportion of variance in saplings (55.8%), adult (51.3%), and all trees (58.3%). In saplings and in all trees, leaflets from different phyllotactic positions showed no distinct separation in PCA space (Figure 3A,C), whereas in adult trees, leaflets at position 6 exhibited a clear positional shift toward the direction of higher SLA and lower LT (Figure 3B).

3.3. Effects of Phyllotaxy and Environmental Factors on Leaf Traits Across Ontogenetic Stages

The influence of phyllotaxy and environmental factors on leaf traits was stronger in saplings than in adults (

Table 2. Generalized linear model (GLM) analysis of the effects of phyllotaxy, soil nitrogen content, soil water content, and canopy openness on leaf economic and stomatal traits across ontogenetic stages.

Ontogenetic Stages	Leaf Trait	Phyllotaxy		Soil Total Nitrogen		Soil Water Content		Canopy Openness		Intercept
		Estimate Std.	p Value	Estimate Std.	p Value	Estimate Std.	p Value	Estimate Std.	p Value	Estimate Std.	p Value
Sapling	SLA	0.028	0.002 **	0.062	<0.001 ***	−0.740	<0.001 ***	−0.012	0.625	2.075	<0.001 ***
	LDMC	−0.013	<0.001 ***	0.013	<0.001 ***	−0.245	<0.001 ***	−0.026	0.002 **	0.638	<0.001 ***
	LN	0.015	<0.001 ***	0.015	<0.001 ***	−0.184	0.005 **	−0.023	0.038	1.610	<0.001 ***
	LT	−0.005	0.432	−0.037	<0.001 ***	0.522	<0.001 ***	0.071	<0.001 ***	−0.026	0.778
	SL	0.003	0.322	−0.004	0.144	0.228	<0.001 ***	0.018	0.048	1.086	<0.001 ***
	SW	0.005	0.257	−0.002	0.612	0.239	0.004 **	0.024	0.091	0.954	<0.001 ***
	SD	−0.008	0.021 *	0.0003	0.911	−0.110	0.056	0.016	0.109	2.161	<0.001 ***
	SPI	−0.001	0.782	−0.008	0.196	0.343	0.003 **	0.052	0.009 **	0.334	<0.001 ***
Adult tree	SLA	0.012	0.051	−0.012	0.015 *	−0.517	<0.001 ***	−0.089	0.014 *	2.545	<0.001 ***
	LDMC	−0.012	<0.001 ***	−0.001	0.481	−0.471	<0.001 ***	−0.087	<0.001 ***	0.906	<0.001 ***
	LN	0.012	<0.001 ***	0.002	0.476	−0.049	0.434	0.021	0.181	1.565	<0.001 ***
	LT	−0.021	<0.001 ***	−0.001	0.756	0.532	<0.001 ***	0.228	<0.001 ***	−0.418	0.004 **
	SL	0.001	0.505	0.0004	0.777	−0.165	<0.001 ***	−0.031	0.004 **	1.415	<0.001 ***
	SW	0.001	0.976	0.001	0.981	−2.097	0.001 **	−2.055	0.186	1.345	<0.001 ***
	SD	0.001	0.753	0.004	0.171	0.276	0.005 **	0.037	0.115	1.808	<0.001 ***
	SPI	0.004	0.301	0.005	0.101	−0.051	0.593	−0.024	0.303	0.636	<0.001 ***

Note: SLA: Specific leaf area, cm²/g; LDMC: Leaf dry matter content, g/g; LN: Leaf nitrogen content, mg/g; LT: Leaf thickness, mm; SL: Stomatal length, μm; SW: Stomatal width, μm; SD: Stomatal density, stoma/mm²; SPI: Stomatal pore index, %. *: p < 0.05; **: p < 0.01; ***: p < 0.001. Estimated std. is the standardized regression coefficient.

). Among all examined factors, SWC showed the strongest associations with leaf traits across both ontogenetic stages, with all traits—except SD in saplings and LN and SPI in adults—being strongly related to SWC (Table 2). Phyllotaxy significantly influenced most leaf economic traits in both saplings and adults but had little effect on stomatal traits, except for SD in saplings (Table 2). STN affected all economic traits in saplings, whereas in adults only SLA was related to STN; no stomatal traits were influenced by STN at either stage (Table 2). CO influenced LDMC, LT, and SPI in saplings, and SLA, LDMC, LT, and SL in adults (Table 2). Moreover, the correlations between traits and environmental factors at the leaflet scale were weaker than those at the compound-leaf scale (Table 2 and

Table A2. Correlations between leaf traits and soil nitrogen content, soil water content and canopy openness across phyllotaxy in saplings and adult trees. Bold values indicate significant correlations at the 0.05 level.

Ontogenetic Stage	Phyllotaxy	Environmental Factors	SLA	LDMC	LT	LN	SL	SW	SD	SPI
Sapling	1	STN	0.32	0.11	−0.58	0.41	−0.31	0.28	−0.22	−0.43
		SWC	−0.22	−0.34	0	0.06	−0.16	−0.08	−0.57	−0.37
		CO	−0.01	0.04	0.23	−0.33	0.16	0.35	0.83	0.47
	2	STN	0.53	0.28	−0.55	0.43	0.26	0.02	−0.24	0.05
		SWC	−0.28	−0.28	−0.08	0.09	0.32	0.28	−0.37	0.01
		CO	0.13	0.04	0.1	−0.43	−0.44	−0.4	0.54	−0.02
	3	STN	0.4	0.39	−0.49	0.41	−0.36	0.17	−0.55	−0.57
		SWC	−0.44	−0.23	0.04	−0.04	0.38	0.53	−0.59	−0.21
		CO	0.26	−0.26	0.19	−0.08	−0.16	−0.3	0.38	0.17
	4	STN	0.43	0.5	−0.62	0.54	−0.03	−0.21	−0.42	−0.33
		SWC	−0.42	−0.15	−0.13	0.11	0.61	0.27	−0.9	0.11
		CO	0.16	−0.32	0.33	−0.2	−0.2	0.08	0.64	0.2
	5	STN	0.5	0.31	−0.5	0.53	0.28	0.26	−0.07	0.28
		SWC	−0.16	−0.21	0.03	0.07	0.79	0.5	0	0.74
		CO	−0.02	−0.15	0.48	−0.21	−0.45	−0.32	0.38	−0.3
	6	STN	0.46	0.74	−0.5	0.55	0.27	−0.06	−0.23	0.16
		SWC	−0.23	0.2	−0.19	0.05	0.62	0.28	−0.58	0.38
		CO	0.01	−0.39	0.63	−0.21	0.1	0.34	0.36	0.32
Adult tree	1	STA	−0.45	−0.07	−0.05	−0.06	0.05	0.01	0.09	0.35
		SWC	−0.01	−0.5	−0.41	−0.36	−0.27	−0.23	0.34	0.13
		CO	−0.05	−0.03	0.79	0.57	−0.16	0.08	0.09	−0.14
	2	STN	−0.38	−0.08	−0.09	0	−0.19	−0.05	0.35	0.35
		SWC	−0.21	−0.62	−0.25	−0.5	−0.42	−0.61	0.35	0.13
		CO	0.1	0.14	0.67	0.55	0.33	0.63	−0.38	−0.14
	3	STN	−0.34	0.02	−0.06	0.3	0.3	0.07	0	0.29
		SWC	−0.33	−0.59	−0.38	−0.37	−0.49	−0.58	0.47	0.02
		CO	0.03	0.17	0.74	0.31	0.27	0.48	−0.28	−0.08
	4	STN	−0.16	0.09	0.1	0.22	0	0.06	0.19	0.22
		SWC	−0.49	−0.49	−0.21	−0.32	−0.19	−0.36	0.24	0.06
		CO	0.12	0.08	0.59	0.28	−0.01	0.01	−0.04	−0.08
	5	STN	−0.06	−0.1	0.03	0.23	0.04	0.06	−0.04	0.01
		SWC	−0.47	−0.34	−0.44	−0.39	−0.5	−0.51	0.31	−0.03
		CO	0.11	−0.07	0.7	0.44	0.03	0.33	−0.45	−0.44
	6	STN	−0.08	0.05	−0.1	0.08	0.16	0.16	0.12	0.23
		SWC	−0.45	−0.35	0.05	−0.34	−0.25	−0.59	0.3	0.08
		CO	0.1	−0.09	0.28	0.27	−0.12	0.23	0.14	−0.04

).

4. Discussion

4.1. Economic Traits Are More Sensitive to Phyllotaxy than Stomatal Traits Across Ontogenetic Stages

Economic traits were more sensitive to phyllotaxy than stomatal traits (Figure 1). Along phyllotaxy, economic traits exhibited significant variation from apical to basal leaflets (Figure 1A–D). Basal leaflets, being closer to the vascular tissues, had greater access to water and nitrogen, resulting in higher LN and lower LDMC. In contrast, apical leaflets, positioned farther from the vascular supply and exposed to stronger irradiance, invested more in structural support, leading to higher LT but lower LN. This pattern suggests a functional division of labor and a trade-off between resource acquisition and structural investment along phyllotaxy. Notably, the pattern of trait variation along phyllotaxy is consistent with patterns reported for leaf traits across environmental gradients, such as canopy light availability or soil nutrient supply [32,33]. For example, upper-canopy leaves generally exhibit greater LT and lower LN, whereas lower-canopy leaves maintain higher LN to compensate for shaded conditions [34,35]. Similarly, in nitrogen-rich soils, plants tend to display increased SLA and LN but reduced LDMC, reflecting a resource-acquisitive strategy. Our results suggest that the resource acquisition–structural investment trade-off described by the leaf economics spectrum extends beyond species and environmental gradients to the leaflet scale within compound leaves.

Stomatal traits exhibited weaker variation along phyllotaxy compared with economic traits, indicating a higher degree of conservatism (Figure 1E–H). In this study, SL, SW, and SPI showed no significant differences across phyllotaxy. Such stability may reflect strong functional constraints on stomatal structure and distribution [36]. By contrast, SD followed a “declining at the apex–lowest in the middle–increasing at the base” pattern, suggesting spatial optimization of stomatal distribution. Apical leaflets, typically exposed to stronger irradiance and higher transpirational demand, maintained higher SD to support carbon assimilation and thermal regulation; middle leaflets, influenced by shading and epidermal cell expansion, exhibited relatively lower SD per unit area; whereas basal leaflets, located near the main vein and receiving more abundant water and nutrient supply, sustained relatively higher SD.

Ontogenetic stage had significant effects on certain traits at a given phyllotactic position (Figure 1). For economic traits, SLA and LDMC were significantly lower in adult trees than in saplings, reflecting a developmental shift in carbon allocation strategies. Reduced SLA implies higher construction costs per unit leaf area, which may enhance leaf tolerance to stresses in the canopy environment, whereas lower LDMC suggests a decline in the proportion of dry matter, potentially associated with higher leaf water content that contributes to functional stability under prolonged light and water fluctuations [2,37]. By contrast, LN did not differ significantly between stages, indicating that chemical traits are less responsive to ontogenetic development. Such ontogenetic independence of chemical traits has also been documented in other temperate tree species, where structural traits tend to be more developmentally plastic, while chemical traits remain relatively conservative [9]. For stomatal traits, adult trees exhibited greater SL and SW across most phyllotactic positions, suggesting that adult trees may maintain gas-exchange capacity by increasing stomatal size. In contrast, saplings showed higher SD only at the first leaflet position, which may reflect a demand for rapid carbon acquisition during the early stages of compound leaf development. Moreover, the SPI was significantly higher in adult trees than in saplings at leaflet positions 2–4, further indicating a stronger capacity of mature individuals to balance carbon assimilation and transpiration. Overall, the relatively weak responses of stomatal traits to both phyllotaxy and ontogenetic stage highlight their stronger functional constraints.

4.2. Clear Separation of Phyllotaxy in Adults but Not in Saplings

Within-group correlations among economic and stomatal traits remained largely consistent across ontogenetic stages, whereas cross-group correlations showed distinct differences (Figure 2). Correlations among leaf economic traits generally conformed to the global leaf economics spectrum, reflecting coordinated trade-offs between resource acquisition and structural investment [38,39,40]. Likewise, coordination among stomatal traits indicates that stomatal pore index is jointly regulated by stomatal size and density [41,42,43,44]. The weakening of correlations between SD, SLA, and LN from saplings to adult trees suggests a developmental decoupling of trait integration. First, as trees grow taller, increased hydraulic path length and resistance impose stronger hydraulic limitations on leaf gas exchange, reducing the tight coordination between stomatal deployment and photosynthetic investment that is typical of juveniles. This mechanism is widely discussed under the hydraulic-limitation framework. Second, ontogeny is associated with anatomical reconfiguration of the lamina, and with changes in nitrogen partitioning within leaves (a higher fraction allocated to structural pools and a lower fraction to photosynthetic machinery), which lowers the coupling between LN and stomatal traits [16,45]. Ontogenetic differentiation in trait coupling mirrors a broader shift from acquisitive to conservative resource-use strategies during plant development. Overall, trait–trait correlations were consistently stronger at the compound-leaf scale than at the leaflet scale in both saplings and adults, highlighting that trait coordination is driven primarily by ontogenetic stage rather than by phyllotaxy.

Although the major axes of trait coordination were consistent across saplings, adult trees, and all trees (with the first two axes together explaining 51%–58% of the variation), the spatial arrangement of traits differed (Figure 3). In saplings and in all trees, leaflets from different phyllotactic positions showed substantial overlap in trait space, suggesting limited positional differentiation at early developmental stages. In contrast, in adult trees, leaflets at position 6 exhibited a clear positional shift toward the direction of higher SLA and lower LT (Figure 3B). This pattern is consistent with the results presented in Figure 1, suggesting that distal leaflets (position 6) tend to adopt a more resource-acquisitive trait strategy under upper-canopy light and vapor-pressure deficit conditions. This ontogeny-dependent differentiation likely reflects increasing within-crown environmental heterogeneity—particularly stronger light gradients and vapor-pressure deficit in the upper canopy—that accentuate structural and physiological adjustments among leaflets along the compound leaf axis. These findings suggest that phyllotaxis-related trait differentiation strengthens with ontogenetic development.

4.3. Stage-Dependent Responses of Leaf Traits to Phyllotaxy and Environmental Factors

Our results demonstrate that the influence of phyllotaxy and environmental factors on leaf traits was generally stronger in saplings than in adult trees, suggesting that early ontogenetic stages are more sensitive to resource availability, whereas adult trees rely on more robust and conservative resource–use strategies to cope with canopy resource heterogeneity (Table 2). SWC was the main driver of trait variation (Table 2). Previous studies have shown that, as SWC increases, turgor-driven cell expansion promotes lamina growth, resulting in higher SLA and generally larger stomata; conversely, declining soil moisture shifts leaf traits toward thicker tissues and higher construction costs (lower SLA) [27,46]. In our study, SWC differentially affected traits across ontogenetic stages: in saplings, SWC significantly affected SLA, SL, SW, and SD, highlighting the critical role of water in constraining leaf construction and stomatal distribution; and in adult trees, SWC significantly influenced nearly all traits except LN and SPI, indicating that water availability is a key driver of trait variation under heterogeneous canopy conditions. Our results are consistent with findings reported by Cavender-Bares and Bazzaz [47] and Peters et al. [48].

In contrast, the effects of STN and CO were more stage-specific. In saplings, STN significantly influenced leaf economic traits, indicating that nitrogen supply primarily drives early-stage variation in carbon acquisition and within-leaf N allocation, a stage dependency consistent with comparative evidence that seedlings (or small trees) are more responsive to nitrogen enrichment and nutrient additions than mature individuals [19,49]. In adult trees, responses to nitrogen addition were weak relative to seedlings—STN affected only SLA—consistent with comparative and long-term field evidence; this attenuation likely reflects more stable internal nitrogen redistribution and homeostatic utilization in mature individuals, which reduces sensitivity to external N inputs [4,19]. CO exerted stronger effects in adult trees: its significant influence on SLA, LDMC, LT, and SL suggests that under distinct canopy light gradients, adult trees enhance both structural robustness and stomatal functioning to sustain photosynthetic balance. By comparison, CO had weaker effects in saplings, probably because their limited stature confines leaves to shaded understory conditions where local differences in canopy openness have little impact on actual light availability. This indicates that sapling responses to light are governed primarily by understory shade-adaptation mechanisms rather than by local canopy structure. In addition, we found that correlations between traits and environmental factors were generally weaker at the leaflet scale than at the compound-leaf scale (Table A2). This indicates that, although microenvironmental heterogeneity within compound leaves can drive localized trait differences, overall trait patterns are more strongly constrained by integrative processes operating at the whole-leaf level. Moreover, future studies could extend the comparison of leaf trait responses to phyllotactic and environmental gradients across different ontogenetic stages to multi-seasonal or multi-year scales. Temporal replication would help determine whether the trait differentiation along phyllotactic and ontogenetic gradients remains consistent under interannual climatic fluctuations, such as variations in temperature, precipitation, or light availability. Such temporal-scale comparisons would provide valuable insights into the adaptive stability of plants under long-term climatic variability.

5. Conclusions

Our study clearly showed that phyllotaxy and environmental factors jointly shaped the variation in leaf economic and stomatal traits in Fraxinus mandshurica, with effects strongly dependent on ontogenetic stage. Economic traits exhibited significant variation along phyllotaxy, indicating differences among leaflets in resource acquisition strategies, whereas stomatal traits were relatively conservative, consistent with their anatomical and functional constraints. Traits in saplings were more strongly shaped by STN and SWC, indicating their dependence on resource supply during early growth, while traits in adult were primarily influenced by SWC and CO, reflecting adaptation to the heterogeneous canopy environment. Moreover, trait–environment associations were weaker at the leaflet scale than at the compound-leaf scale, demonstrating clear scale dependence. Overall, our findings extend the resource acquisition–structural investment trade-off described by the global leaf economics spectrum to the leaflet level within compound leaves, and highlight the key role of ontogenetic stage in regulating trait variation and trait–environment relationships.