Root Foraging Behavior of Two Agronomical Herbs Subjected to Heterogeneous P Pattern and High Ca Stress

Abstract

Ecosystems are vulnerable to large areas of rocky desertification, which results in patchy soils and stone-inlaid soils. Karst landforms are typically characterized by heterogeneous phosphorus (P) distributions in soils at high calcium (Ca), but root foraging behavior has not been fully documented in agronomical plants. In this study, Bidens pilosa L. and Plantago asiatica L. were raised in pots in a simulated soil environment with sands at high Ca (2 g kg⁻¹) and low Ca (0.63 g kg⁻¹) levels. Inner spaces were divided into four sections to receive P in homogeneous (Homo.) (four quarters: 2 mg P kg⁻¹) or heterogenous (Hete.) (one quarter: 8 mg P kg⁻¹; three quarters: no-P input) patterns. Both species had longer roots in high P sections compared to no P sections. Foraging scale (highest length or surface-area(SA)) was higher in P. asiatica plants subjected to the Hete. pattern than to the Homo. pattern in low Ca pots. Foraging precision (length or SA differences between P patches as a proportion of the total) was also higher for P. asiatica subjected to the Hete. pattern but did not change in response to Ca level or P placement pattern. Overall, P. asiatica has a higher foraging ability than B. pilosa because of higher levels of foraging scale and precision from high-P (8 mg kg⁻¹) patches in soils subjected to low Ca (0.63 g kg⁻¹).

1. Introduction

Karst ecosystems account for 12% of the world’s total land area [1]. Karst topography makes up 20–25% of the ice-free land surface on Earth [2], which is characterized by rocky desertification (RD) [3,4]. It is an irreversible ecosystem failure with abundant carbonate rocks like limestone, dolomite, and marble [4], making rocky landforms one of the most fragile terrestrial ecosystems in the world. The driving factor of RD, at least partly, is the high level of soil calcium (Ca) concentration [1,4]. High soil Ca is a predictor of vegetation degradation in RD and is associated with soil bacterial network [4,5], Ca hyperaccumulation in plant roots [1], and arrests of soil organic carbon [6]. As a result, local plant populations are naturally selected by their adaptability to high Ca conditions [1,7]. More information on the response of belowground karst plant parts to high soil Ca is needed to guide rehabilitation.

Phosphorus (P) is one of the major limiting factors for the growth of many plants. Low P availability is related to long-term rock weathering [8] and low mobility [9]. In ecosystems with RD, plants in calcareous soils are frequently subjected to low P stress [10,11]. In degraded karst ecosystems with RD, low P stress is strengthened in calcareous soils because P is strongly adsorbed and precipitated by Ca [12]. RD was reported to decrease total P and available P concentrations in karst soils [13]. Soil P availability was also reported to be unaffected by RD in a karst valley [14]. Therefore, it is the high levels of Ca in soil that limit P acquisition by local plants. Given that P diffuses very slowly in soils [15], with diffusion coefficients of 10⁻⁹ to 10⁻⁷ cm² s⁻¹ [16], its mobility is further limited by Ca precipitation [17]. We need more evidence to demonstrate the impact of high Ca levels on P acquisition by plant roots that are acclimated to RD.

P acquisition can also be achieved by the movement of plant roots toward soil nutrients. The placement of fine roots for acquiring soil nutrients is termed as root foraging behavior [18]. Low mobility of P compounds in soils contributes to a basic heterogeneity of P placement in soils [19]. Plants can modify root architecture to place more fine roots in P-enriched soil patches at the lowest cost of root morphology [20,21,22,23]. In calcareous soils, the precipitation of P further increases the heterogeneity of P patches; hence, root foraging behavior in karst plants has uncertainties for P acquisition. A root-split experiment is a classic instrument to assess plant root foraging of soil P when facing various soil heterogeneities [21,22]. To our knowledge, studies on root foraging behavior for P acquisition by karst plants are scant. Some field investigations confused “so-called” root foraging ability in karst soils, which was evaluated by measuring the gross response of whole-root morphology in different types of karst soils [24,25]. Research on fine root placement in heterogeneous soil patches in a root-split experiment is highly scarce for karst plants. Information on combined high-Ca and low-P effects on the root foraging behavior of plants that dwell in karst habitats is even more limited.

Root foraging, an ability to capture resources, can be described by foraging-scale and -precision in competition by plants for resources [26]. A foraging scale was used to describe the extent to which dominant plants in a community monopolize mineral nutrients in soils. Foraging precision was used to quantify the possibility and speed of subordinate plants to place their fine roots in nutrient enriched patches of soils. These two variables were frequently tested in root-split experiments and further synthesized to evaluate the ability to capture nutrients by plants with different root sizes [27]. That is, plants with a larger root system (scale) have a higher possibility to seize soil nutrients in a broad range, but at less selection in placing roots in nutrient-rich patches. In contrast, small-root plants have a higher flexibility and precision to place roots in heterogeneous soil patches. Therefore, there may be a trade-off between the spatial scale over which plants forage and the precision with which they are able to capture nutrients with fine root proliferation [28]. Both scale and precision are usually used in combination to evaluate different aspects of root foraging behavior [21,22,23]. Root foraging behavior has a higher complexity for plants acclimated to karst soils. High Ca in karst soils can directly generate a stress that may induce a greater allocation of dry mass to roots [29] but may also influence root tissue density [30]. The response of root size to Ca stress increases uncertainties in foraging scale. Precipitation of P delays movement and increases heterogeneous distribution, which may impact foraging precision. More work is essentially needed to detect the direct responses of root foraging behavior in karst plants. This is an important precondition depending on which strategies are used to cope with RD during ecosystem rehabilitation.

Bidens pilosa L. was introduced from tropical America to pan-tropic regions [31]. Due to a strong acclimation to varied tropical conditions, Bidens pilosa has been listed as a noxious invasive weed [32,33]. It dominates herb communities in karst soils and has been found to place fine roots in heterogeneous soil environments on karst landforms [34,35]. Compared to native species, its root foraging behavior was characterized by a higher scale but a lower precision [36]. Plantago asiatica L. is a perennial forb from the Plantaginaceae family with colonization in shady habitats [37,38]. Plantago asiatica can dwell in karst rock hill areas as scattered shade-tolerant individuals [39]. It has a relatively larger root system than local legume species [37], but this did not confer any priority in foraging scale. Plantago asiatica has a similar foraging precision compared to local deciduous tree species, both of which are easily interrupted by herbivores [40]. Both B. pilosa and P. asiatica roots had high foraging scale potential in karst soils, but their root morphology may be reduced by high Ca. Low foraging precision may be further decreased by a higher level of soil P heterogeneity. Therefore, these two plants are desired karst dwellers that may proliferate long roots to seize P but be slowed by high Ca.

In this study, B. pilosa and P. asiatica seedlings were cultured in a classic root-split experimental model. Soil characteristics simulated ecosystems with high Ca and heterogeneous P placement. Our objective was to test the scale and precision of root foraging behavior in the two plant species subjected to heterogeneous P supply at high Ca condition. We also aimed to detect the trade-off between root foraging scale and precision and their relationships with fine root morphology in combined Ca and P environments. We hypothesized that (i) both species placed more fine roots in P enriched patches, but (ii) both root foraging scale and precision would be limited in soils subjected to high Ca. Our study will fill the knowledge gap on karst plants’ root foraging behavior in heterogeneous P placement at high Ca. Our results can supply empirical evidence about foraging scale and precision in karst plants.

2. Materials and Methods

2.1. Plant Materials and Experiment Commencement

Seeds of Bidens Pilosa and Plantago asiatica were obtained from Guiyang Center Nursery (26°32′ N, 106°40′ E), Guiyang, Guizhou, China. Seeds were collected from a region with typical RD landforms in central Guizhou (26°13′‒26°32′ N, 106°07′‒106°23′ E). In late March 2018, seeds were soaked in distilled water for 8 h at indoor temperature and sterilized by being soaked in 5% potassium permanganate (w/w) for 12 h [23]. Sterilized seeds were soaked in water again to remove residual chemicals. Cleaned seeds were sown into mashed peat for germination at a constant temperature of 36 °C and a relative humidity of 85% [41]. When seeds germinated to the stage when the first pair of cotyledons burst, they were transplanted to pots (top Ø 21 cm, bottom Ø 15.5 cm, height 15.0 cm) filled with washed sands.

2.2. Experiment Design and Arrangement

The experiment was arranged in a split-plot design. The main plot was the simulated high Ca level of karst soils, while low Ca levels served as the control. The subplot was two contrasting patterns of P supply. The high Ca level was employed to be 2 g kg⁻¹, and the low Ca level to be 0.63 g kg⁻¹ [1]. These two levels were derived from averaged contents of exchangeable Ca²⁺ in lands with intensive and low RD, respectively [42]. Averages of exchangeable Ca²⁺ were calculated using data of soil samples at the non-rocky side and the rocky side. P was supplied in two types of placements, as a homogeneous pattern and a heterogeneous pattern (Figure 1). In both patterns, the inner space of a planting pot was divided into four sections with an even volume. In the homogeneous pattern, all four sections received P supplies at a common rate of 2 mg kg⁻¹; in the heterogeneous pattern, only one random section received P supply at a rate of 8 mg kg⁻¹, leaving the other three sections with no P input. The rate of P supply, at 2 mg kg⁻¹, was derived from available P content in shallow and rocky residual soils at the bedrock layer [25]. In a karst ecosystem, soil P content is highly heterogeneous in patches at bedrock layer because of the direct delivery of P from rock weathering and the frequent proliferation and penetration of plant roots for foraging [43]. The high rate of 8 mg P kg⁻¹ was the sum of four low P rates in a pot, which is equal to the average of available P content in soils at a foraging range between rhizosphere and surrounding bulks [24].

Sands were filled to a depth of 2 cm below the top of a pot. Plastic sheets (13 cm in height) were inserted into pots to divide the inner space into four sections of equal volume. Plants were transplanted in the center and then Ca and P were applied once a week. Ca²⁺ ions were applied as calcium chloride (CaCl₂) at an even amount to all four sections of a pot. Subsequently, P was applied as phosphoric acid (H₃PO₄). Both Ca and P were applied as solutions with pH adjusted to 5.6 [1] using sodium hydroxide and hydrochloric acid. In the treatment of the heterogeneous pattern, pot sections without P supply just received distilled water at the same volume as the 8 mg P kg⁻¹ solution. Both Ca and P were delivered using 5 mL pipettors to the surface of sands at two different spots to avoid contamination of plant shoots [44]. All pots with combined Ca and P treatments were arranged in a block, which was replicated ten times with random placements after treatment implement every week. All Ca and P solutions were applied to the surface of sands without contamination to the surface of plant leaves [21]. All pots were watered twice or three times a week depending on local weather conditions. Plants were watered to pot capacity in a common volume by measuring the weight of other sand-filled pots without plant culture. All pots were watered by mild flows of water to avoid P loss in leaching.

2.3. Plant Culture and Sampling

Seedlings were cultured in a Guizhou University greenhouse. Fans and curtains were used to control the daytime temperature to be no higher than 42 °C in summer. Artificial lighting (Pudao photoelectricity, Zhiluntuowei A&F S&T Inc., Changchun, China) was used to maintain photosynthetic photon flux density (PPFD) to a range of 70–77 µmol m⁻² s⁻¹ during daytime [45,46]. To benefit root growth, we used a spectrum of 40% red light, 58% green light, and 2% blue light, which was proven to promote fine root morphology [47]. Seedlings were fed nitrogen (N) and micro-element nutrients every week with the application of CaCl₂ solutions. Nutrient solutions included 4 mM ammonium nitrate, 0.5 mM potassium chloride, 0.6 mM magnesium sulfate, 20 µM ferric chloride, 6.0 µM manganese chloride, 16 µM boric acid, 0.3 µM zinc chloride, 0.3 µM cupric chloride, and 0.3 µM sodium molybdate. This solution has been used to culture transplanted plants and benefits newly grown root morphology [48,49]. Chloride chemicals were commonly used as nutrients in studies on plant physiology [50,51].

Four months after the experiment commencement, all seedling roots were carefully excised from shoots at the soil surface. Roots were further divided into four sections from the tap root axis. All roots were immediately rinsed in distilled water and placed in moist towels until transported to the laboratory. During the whole sampling process, roots were kept as an intact system for every pot section. Fine roots with a diameter less than 1 mm were kept as research objectives. All root chains were laid on the operational panel of the scanner (HP Deskjet 1510, HP Inc., Palo Alto, CA, USA). Tweezers were used to carefully separate every lateral root from overlap to obtain the most accurate evaluation on fine root tip number and fork number. Fine root morphology per pot section was scanned to generate a projected image at a resolution over 110 pixels cm⁻¹ [22,52]. WinRHIZO software (Regent Instrument Inc., Calgary, Canada) was used to analyze fine root morphologies in length, surface area, tip number, and fork number.

2.4. Calculation and Statistical Analysis

Root foraging scale is widely accepted as a term of a rhizopheric range that root morphology can monopolize [26]. Based on this term, fine root morphology in soil patches was used to evaluate foraging scale [22,53], but average value cannot fully reflect the maximum range of root monopolization. In this study, we employed the highest record of data, about fine root morphology in a pot section, as the parameter to assess root foraging scale [54]. In pots with homogeneous P supply, foraging scale was calculated as the average of the highest records of fine root morphology (length or surface area) of the four sections; in pots with heterogeneous P supply, foraging scale was calculated as the average of fine root morphology in the three sections with no P input or as the highest record in the P enriched section.

Root foraging precision was assessed by expressing the fine root placement in the nutrient enriched patch as a proportion of the total roots [26,27]. Regarding how plants place fine roots in a random rhizospheric orientation, fine roots can also be placed in nutrient poor patches in the heterogeneous soil environment, which impacts foraging precision and should be excluded from calculation. Therefore, in this study we used an upgraded equation to assess foraging precision (${\mathrm{RF}}_{\mathrm{precision}}$):

$${\mathrm{RF}}_{\mathrm{precision}}=\frac{{\mathrm{RF}}_{\mathrm{sensitivity}}}{{\mathrm{Root}}_{\mathrm{total}}}\times 100\%$$

(1)

where ${\mathrm{RF}}_{\mathrm{sensitivity}}$ is the difference in the fine root placement of two random soil patches [27]. In homogeneous P supply, ${\mathrm{RF}}_{\mathrm{sensitivity}}$ is the fine root difference in morphology (length or surface area) of random two sections [21]; in heterogeneous P supply, ${\mathrm{RF}}_{\mathrm{sensitivity}}$ is the fine root difference in morphology between the P enriched section and a random no-P section [22,23]. ${\mathrm{Root}}_{\mathrm{total}}$ is the total amount of fine root morphology (length or surface area) in a pot.

All data were analyzed and calculated using SAS v.9.4 software (SAS Inst., New York, NY, USA). All of our data passed tests of normality and homogeneity of variance; hence, no transformation was used. Analysis of variance (ANOVA) was used to analyze fine root foraging behavior in two sets of statistical methods. Fine root morphology (length, surface area, tip number, and fork number) was analyzed as a two-way ANOVA model with Ca level across all four sections (low: 0.63 g kg⁻¹, high: 2 g kg⁻¹), P supply level per pot section (no addition, low: 2 mg kg⁻¹, high: 8 mg kg⁻¹), and their interaction as independent variables. Fine root foraging behavior (scale, sensitivity, and precision) was evaluated by another two-way ANOVA model with Ca level, P placement pattern (heterogeneity vs. homogeneity), and their interaction as independent variables. The basic combined treatment (Ca × P) per pot was replicated 10 times. When significant effects were detected, results were compared by Duncan’s new multiple range test at the p < 0.05 level to cope with comparisons of means with unequal numbers of replications [55]. The Pearson correlation was used to detect relationships between pairs of variables from root foraging scale, sensitivity, and precision. Multivariate linear regression was used to detect the combined contributions of fine root morphologies to root foraging behavior variables.

3. Results

3.1. Fine Root Morphology Per Pot Section

Substrate Ca level and P dose had an interactive effect on fine root length and number of forks in B. pilosa plants, but this interaction generated significant effects on all four types of fine root morphologies in P. asiatica plants (

Table 1. Analysis of variance (ANOVA) of calcium (Ca) and phosphorus (P) concentrations and their interaction (Ca × P) on fine root parameters of individuals in Bidens pilosa L. and Plantago asiatica L. seedlings.

Fine Root Parameter	Source of Variance
	Ca	P	Ca × P
Bidens pilosa
Length	5.34 *	16.12 ***	3.38 *
Surface area	0.99	9.07	0.17
Number of tips	4.43 *	7.66 ***	2.37
Number of forks	3.82	8.92 ***	3.94 *
Plantago asiatica
Length	228.54 ***	51.14 ***	20.25 ***
Surface area	129.5 ***	35.43 ***	14.54 ***
Number of tips	132.05 ***	39.98 ***	11.10 ***
Number of forks	132.81 ***	41.04 ***	17.21 ***

Note: Number of asterisks remarks range of p values of significant responses: *, p < 0.05; ***, p < 0.001.

).

Fine root length was higher in the section with high P supply and low Ca level than that in pots with high Ca level across all three P treatments for both B. pilosa and P. asiatica plants (Figure 2A,B). Fine root length was also greater in sections with high P supply compared to those in sections with no P supply in pots with both Ca levels for both plant species.

In B. pilosa, fine root surface area was 14.95% lower in pots with high Ca level (721.08 ± 300.77 cm²; mean ± standard error unit, the same below) than that in pots with low Ca level (847.81 ± 314.95 cm²) (Figure 2C). Fine root surface area of B. pilosa plants showed an increasing trend with the dose of P supply (no P, 624.39 ± 224.93 cm²; low P, 820.91 ± 243.09 cm²; high P, 1118.81 ± 567.42 cm²) (Figure 2C). In P. asiatica, fine root surface area also increased with the dose of P supply in sections of pots with high and low Ca levels (Figure 2D).

In B. pilosa plants, high Ca level lowered fine root tip number (169.07 ± 65.31 individuals), compared to low Ca level (208.61 ± 125.63 individuals) by 18.95% (Figure 2E). High P supply resulted in higher fine root tip number (269.59 ± 212.97 individuals) compared to those in low P supply (196.27 ± 75.90 individuals) and no P supply (152.02 ± 69.80 individuals) (Figure 2E). In P. asiatica plants, high P supply increased fine root tip number compared to the no-P supply treatment in pots with both high and low Ca levels (Figure 2F).

In B. pilosa plants, root fork number was higher in sections receiving high P supply than in sections with no P in pots with low Ca level, but root fork number was not changed by P supply in pots with high Ca level (Figure 2G). In P. asiatica, high P supply increased root fork compared to the no-P supply treatment in pots with both high and low Ca levels (Figure 2H).

3.2. Root Foraging Scale

Roots of B. pilosa plants did not show any response in foraging scale in length or surface area to substrate Ca level or P placement pattern (

Table 2. Analysis of variance (ANOVA) of Ca concentration, P placement pattern, and their interaction (Ca × Pattern) on root foraging sensitivity, precision, and scale of individuals in Bidens pilosa L. and Plantago asiatica L.

Root Foraging Parameter	Source of Variance
	Ca	Pattern	Ca × Pattern
Bidens pilosa
Scale of root length	2.36	1.00	2.42
Scale of root surface area	0.58	2.82	0.01
Sensitivity of root length	0.83	8.85 **	1.46
Sensitivity of root surface area	0.16	1.27	0.01
Precision of root length	0.01	12.93 ***	0.18
Precision of root surface area	0.01	2.26	0.25
Plantago asiatica
Scale of root length	69.70 ***	11.90 **	10.65 **
Scale of root surface area	39.74 ***	4.13 *	4.79 *
Sensitivity of root length	8.02 **	17.56 ***	1.03
Sensitivity of root surface area	13.60 ***	7.47 *	0.97
Precision of root length	0.90	31.97 ***	0.01
Precision of root surface area	0.03	23.83 ***	0.03

Note: Number of asterisks remarks range of p values of significant responses: *, p < 0.05; **, p < 0.01; ***, p < 0.001.

). Observed root foraging scale of length was averaged to range between 799.62 ± 315.26 and 1119.16 ± 335.13 cm, and scale of surface area ranged between 221.45 ± 107.55 and 313.00 ± 109.91 cm² (Figure 3A,C).

Instead, substrate Ca and P supply had a combined effect on root foraging scale in P. asiatica plants. In pots with low Ca levels, the heterogeneous P placement pattern resulted in a higher foraging scale of root length compared to the homogeneous pattern (Figure 3B), but the effect of P placement pattern disappeared on the foraging scale of root surface area (Figure 3D). Root foraging scale was generally higher in pots with low Ca level than with high Ca level, gauged through length (low, 924.24 ± 341.77 cm; high, 104.54 ± 70.61 cm) or surface area (low, 128.31 ± 66.22 cm²; high, 40.90 ± 14.45 cm²).

3.3. Foot Foraging Sensitivity

In B. pilosa plants, the P placement pattern had a significant main effect on the root foraging sensitivity of root length (Table 2). Compared to P placement in the heterogeneous pattern, that in the homogeneous pattern reduced root foraging sensitivity of length by 65.18% (heterogeneity, 602.23 ± 499.03 cm; homogeneity, 209.67 ± 104.87 cm) (Figure 4C). Both substrate Ca level and P placement had significant effects on the root foraging sensitivity of length and surface area (Table 2). Root foraging sensitivity of length and surface area were both higher in pots with a low Ca level than those with a high Ca level (Figure 4B,F). The heterogeneous pattern of P placement increased root foraging length and surface area relative to the homogeneous pattern (Figure 4D,H).

3.4. Foot Foraging Precision

P placement pattern showed a significant main effect on root foraging precision of length in B. pilosa plants (Table 2). Compared to foraging precision of length in homogeneous pattern (6.72 ± 3.80 cm), that in the heterogeneous pattern (19.25 ± 12.53 cm) increased by 186.52%. In P. asiatica plants, the heterogeneous pattern also resulted in higher root foraging precision of length and surface area compared to the homogeneous pattern (Figure 5D,H).

3.5. Regressions of Root Foraging Parameters against Fine Root Morphologies

In B. pilosa plants, fine root length was estimated to positively contribute to root foraging scale of length (

Table 3. Multivariate linear regression of root foraging parameters of scale, sensitivity, and precision of length and surface area against dependent variables of fine root morphologies of length, surface area, tip number, and fork number.

Root Foraging	Root Morphology	Length			Surface Area
		Estimate	F Value	p Value	Estimate	F Value	p Value
	Bidens pilosa
Scale	Intercept	−24.14	0.07	0.7925	−3.53	0.02	0.8802
	Length	1.33	141.22	<0.0001
	Surface area				1.11	26.56	<0.0001
	Tip number
	Fork number
Sensitivity	Intercept	−28.34	0.03	0.86	−22.17	0.28	0.6006
	Length
	Surface area
	Tip number
	Fork number	2.3	8.53	0.0058	0.78	14.74	0.0005
Precision	Intercept
	Length
	Surface area
	Tip number
	Fork number
	Plantago asiatica
Scale	Intercept	42.59	2.64	0.113	99.69	9.56	0.0039
	Length	0.63	13.77	0.0007	−0.18	17.08	0.0002
	Surface area	5.15	17.96	0.0001	3.15	87.62	<0.0001
	Tip number				−291.42	9.2	0.0045
	Fork number
Sensitivity	Intercept	−698.46	4.82	0.0343	−280.88	23.79	<0.0001
	Length
	Surface area
	Tip number	2747.15	9.54	0.0037	935.26	33.75	<0.0001
	Fork number
Precision	Intercept	25.25	35.87	<0.0001
	Length
	Surface area
	Tip number
	Fork number	−0.06	6.27	0.0167

Note: All values reported in the table are significant estimates at 0.05 level.

). Fine root surface area also had a positive contribution to the root foraging scale of surface area. The fork number of fine roots was estimated to positively contribute to root foraging sensitivity of both length and surface area. No fine root morphological parameters had any estimated contributions to root foraging precision of either length or surface area.

In P. asiatica plants, both fine root length and surface area were estimated to positively contribute to root foraging scale of root length (Table 3). Fine root length and tip number were estimated to negatively contribute to root foraging scale of surface area, but fine root surface area had a positive contribution. Fine root tip number was estimated to positively contribute to root foraging sensitivity of length and surface area. In addition, fine root fork number negatively contributed to root foraging precision of length.

3.6. Relationship between Root Foraging Scale and Precision

In B. pilosa plants, root foraging scale had a positive relationship with the sensitivity of both root length and surface area (

Table 4. Pearson correlation between pairs of parameters assessing root foraging behavior among scale, sensitivity, and precision in Bidens pilosa L. and Plantago asiatica L.

Foraging Parameter	Correlation Coefficients	Scale	Sensitivity	Precision	Scale	Sensitivity	Precision
		Root Length			Root Surface Area
Scale	R	1	0.5317	0.2225	1	0.5832	0.2626
	p	-	0.0004	0.1676	-	<0.0001	0.1017
Sensitivity	R	0.8532	1	0.8797	0.2665	1	0.8656
	p	<0.0001	-	<0.0001	0.0965	-	<0.0001
Precision	R	0.2626	0.8656	1	−0.2258	0.7872	1
	p	0.1017	<0.0001	-	0.1612	<0.0001	-

Note: R and p values in white-color background indicate correlations for Bidens Pilosa and those in italic font indicate correlations for Plantago asiatica.

). In addition, root foraging scale also had a positive relationship with the sensitivity of root length for P. asiatica plants. Root foraging sensitivity had a positive relationship with the precision of both root length and surface area for both plant species. Root foraging scale had no relationship with precision.

4. Discussion

4.1. Fine Root Placement in Heterogenous P Patches in High Ca Soils

Fine root length was generally higher in P-rich sections than in those with no P input. The proliferation of fine roots in P-rich patches of soils was also found as a root foraging behavior in some other plants, such as maize (Zea mays L.), bean (Vicia faba L.) [56], reed (Neyraudia reynaudiana Kunth) [57], and poplar (Populus cathayana Rehder) [58]. The high rate of P supply resulted in higher fine root length relative to the no-P treatment for both plant species. However, fine root placement of the two species differed for root parameters of surface area and tip number. High P supply increased fine root surface area and tip number for P. asiatica plants, but did not impact fine root parameters for B. pilosa plants. In addition, for P. asiatica subjected to high Ca, the high P section resulted in more fine root forks relative to the no-P treatment, but for B. pilosa plants, fine root fork number was unchanged by P supply. Therefore, P. asiatica showed a stronger species-specific root-foraging ability to place fine roots in P-rich patches with a larger surface area, shaped by promoted ramification and new root egress. However, these attributes did not appear in B. pilosa, especially under the high Ca condition. Our high P supply rate of 8 mg kg⁻¹ was adapted from the available P concentration in edaphic spaces for foraging. B. pilosa has the ability to proliferate fine roots for foraging in these spaces [35], but its ability to enlarge its surface area for P acquisition by initiating new roots and ramifying at forks is weak. Species-specific root foraging behavior and increasing new root tips were also reported in tree seedlings of Podocarpus macrophyllus (Thunb.) Sweet and Taxus cuspidate Siebold & Zucc. [22]. The species-specific difference of root surface area was also found in a study involving bamboo (Phyllostachys edulis [Carrière] J.Houz.), Chinese fir (Cunninghamia lanceolata (Lamb.) Hook.), and pine (Pinus massoniana Lamb.) seedlings [59]. Thus, we cannot fully accept our first hypothesis because B. pilosa did not place fine roots as plentiful as P. asiatica did in P-rich patches.

The low rate of 2 mg P kg⁻¹ was adapted from a simulated concentration of available P in shallow soils at the bedrock layer. At high Ca level, sections with low P input did not show increases in most of the fine root morphological parameters for both plant species. At low Ca level, however, low P supply increased fine root morphology in P. asiatica, but it failed to induce any changes in fine roots in B. pilosa. Together, these findings suggest that, in shallow soils at the bedrock layer, high soil Ca restricted root foraging for P acquisition for both species, while low soil Ca restricted root foraging only in B. pilosa plants. It was reported that B. pilosa can accumulate a large amount of Ca in leaves, but its growth and distribution were both very sensitive to soil Ca concentration [60,61]. Our results further revealed that low P acquisition ability is a driving force that restricts the growth of B. pilosa in calcareous soils. P. asiatica’s ability to absorb and accumulate Ca is weaker compared to other species in a same community [62]. This is determined by the weak autonomy of Ca absorption in B. pilosa.

Overall, it can be concluded that both species can proliferate fine roots in sections with high P availability (8 mg kg⁻¹), but B. pilosa cannot increase surface area and tip number of its fine roots. High levels of soil Ca can restrict P foraging by the roots of B. pilosa.

4.2. Root Foraging Behavior and Driving Forces from Fine Root Morphology

Fine root foraging scale was not affected by P placement pattern and soil Ca level for B. pilosa plants. In this species, foraging scale was driven by the change in length and surface area of fine roots, which were also not changed by combined Ca × P. Foraging scale of root length was higher in the heterogeneous P pattern than in the homogeneous pattern for P. asiatica plants subjected to low Ca. It was fine root length and surface area that together contributed to the increase of foraging scale of root length in P. asiatica plants, and surface area had a greater estimated contribution. Therefore, the higher foraging scale in the heterogeneous pattern resulted from a great driving force of surface area. Higher scale of foraging for P is in agreement to the findings in maize [63], bamboo, and Chinese fir [59]. However, when foraging scale was gauged by surface area in P. asiatica plants subjected to low Ca condition, the difference between homogeneous and heterogenous patterns disappeared. Although fine root surface area still had a positive contribution to scale, both length and tip number in fine roots had negative contributions to foraging scale of surface area, and tip number showed a much stronger negative driving force. Therefore, the positive contribution of surface area in fine roots was offset by the negative contributions of length and tip number, which removed the difference between patterns.

Bidens pilosa plants showed no responses in foraging sensitivity and precision to Ca level treatment. This can be explained by the lack of fork-number response to the Ca level, which was predicted to be the unique driver of foraging sensitivity. The low foraging sensitivity of B. pilosa was adapted by the foraging strategy to proliferate roots, along with ramification, in calcareous soils. A high level of Ca was found to exist in the humus layer of contaminated soils where root ramification was impaired [64]. When soil environment was improved to benefit foraging by ramifying root crossings, Ca uptake did not accord [65]. Instead, root foraging sensitivity was lowered by the high Ca level in P. asiatica, which was driven by fine root tip number. Initiation of new roots is a precondition to grow root hair, which is directed by rhizospheric Ca concentration [66]. Higher foraging sensitivity and precision in the heterogeneous pattern concur with findings about fine root difference between contrasting soil patches [21,22,23]. Overall, the contrasting root foraging behaviors between the two plant species result from their strategies to proliferate length and surface area. The strategy of increasing root tips accounts for higher ability in P. asiatica, and the low ability in B. pilosa was due to its dependance on root ramification. Thus, our second hypothesis may be accepted.

4.3. Relationship between Root Foraging Scale and Precision

The trade-off between root foraging scale and precision was put forward by Campbell et al. [26], who hypothesized that foraging scale and precision would be negatively correlated because they assessed abilities to compete resources underground by dominant and subordinate plants in a community, respectively. This hypothesis was confirmed by Grime [28] in a restoring perspective and Chen et al. [36] on pairs of invasive-native plants. However, reviews using synthesized data across studies suggest that this hypothesis did not exist at all [54,67].

In the original study, Campbell et al. [26] exposed his consideration of the concept of foraging scale as the range that a dominant plant can capture mineral nutrients through the development of extensive root systems. However, authors did not give a specific calculation of foraging scale. Thereafter, Wijesinghe et al. [27] provided a clear definition of foraging scale, which is the size of a growing root system. Further studies introduced more calculations to quantify foraging scale, which included parameters of root mass density, root length, root biomass in a patch, and even plant height and relative shoot growth rate [54,67]. Root foraging precision was also defined in various ways, including but not limited to root grow rate, root biomass response, root system size, or even the study duration and plant height [54]. Therefore, the over-varied definitions of foraging scale and definition resulted in contrasting results. Few current studies have employed parameters to assess foraging scale and precision in a meaningful way.

In our study, we did not find any trade-off between foraging scale and precision either. Our parameters retained the original meaning of these two parameters put forward by Campbell et al. [26]. We employed the average of the highest root morphology as the assessment of scale, which reflects the range of root foraging by placing fine roots in heterogeneous soil patches. Other parameters using biomass as a coefficient cannot precisely describe the spatial extent of foraging. There are also some parameters that employ shoot growth and growing period as coefficients. It is unlikely that these parameters show essential relations to root growth extension. Our precision was calculated by the morphological difference from a total root system, which reflects foraging root area and allocation from the total system. Some studies, such as Wijesinghe et al. [27], used biomass difference between heterogeneous patches divided by total root mass as the gauge to assess precision. This is an incorrect expression of the meaning of foraging precision because biomass was invested to the soil patch where root system had dwelled. In addition, root biomass showed different responses to heterogeneous soil patch in comparison with root morphology [21,22].

However, we found a positive relationship between foraging scale and sensitivity. Our sensitivity was calculated as the difference of fine root morphologies between contrasting patches in heterogeneity for the two plant species. This definition was used to assess foraging precision as well [22,68]. Therefore, we conclude that our results did not reveal a trade-off between foraging scale and precision.

5. Conclusions

In this study, B. pilosa and P. asiatica were raised as model dwellers in calcareous soils subjected to P deficit. We found that B. pilosa had a generally lower ability to forage P than P. asiatica. Both species proliferated fine roots to P-rich patches using different strategies. Plantago asiatica enlarged the surface area of fine roots by increasing root initiation and ramification, while B. pilosa increased fine root length. High Ca level formed a stress by limiting foraging scale. Overall, to cope with ecological degradation by RD, it is suggested to use an agronomical strategy to plant P. asiatica in P-rich soils under low-Ca stress. Root foraging behavior can promote fine root proliferation into calcareous soils when the whole root system has already grown to a large scale.