Glyphosate Hormesis Improves Agronomic Characteristics and Yield of Glyphosate-Resistant Soybean Under Field Conditions

Abstract

Brazil, the world’s largest soybean producer, owes its success to the cultivation of glyphosate-resistant (GR) cultivars. However, the soybean yields lag behind those obtained in areas managed for high productivity. Glyphosate-induced hormesis holds promise for increasing crop yields, but the potential evolution of resistance in certain weed species poses a challenge to foliar applications under field conditions. This study assessed the effects of a hormesis-inducing glyphosate dose [90 g acid equivalent (ae) ha⁻¹] on the agronomic characteristics and yield of four GR soybean cultivars. The evaluation was conducted in field settings across various Brazilian locations, considering foliar, seed, and seed + foliar treatments. The results showed variations in dry mass, root nodules, nutrient composition, plant height, pods, and yield, primarily influenced by environmental conditions, soil quality, and, ultimately, the interaction between GR cultivars and treatments. Total dry mass consistently increased with glyphosate, with seed and seed + foliar treatments showing the most substantial increases (7–21%). All three treatments increased nodulation by up to 36% across locations and cultivars, with seed + foliar treatment causing notable increases in nodule dry mass (up to 56%), followed by seed treatment (41%). Nutrient composition, especially for N, P, Br, and Fe, displayed location-dependent variations. Plant height varied among locations and cultivars, with minimal differences between treatments. Glyphosate treatments increased pod numbers (10 to 35%) and yields (11 to 42%) of soybean in seed and seed + foliar treatments. The findings highlight the potential of glyphosate hormesis as a viable tool for improving yields of GR soybean cultivars at the field level. However, the extent of benefits depends on the agronomic conditions of location, choice of cultivars, and herbicide application method.

1. Introduction

Soybean is the main oilseed in the world, cultivated globally in various regions. In 2023, its cultivation occupied approximately 136 million hectares, resulting in a total production of 398,882 million tons [1]. Brazil is the world’s largest producer and exporter of soybeans, contributing 41.9% to global production over 44 million hectares, of which 60.7% is exported. The average soybean yield in the 2022/23 harvest reached 3508 kg ha⁻¹ in the country [2]; however, in areas managed for high yields, averages have exceeded 8 tons ha⁻¹ [3]. This demonstrates that soybean production has significant potential to increase the productivity of this oilseed in Brazil.

Since 1996, genetically modified (GM) crops have been rapidly adopted and expanded across a vast acreage. In 2014, glyphosate-resistant (GR) soybean cultivars already accounted for more than 50% of all herbicide-resistant GM crops, representing over 80% of global soybean cultivation [4]. In Brazil, more than 90% of the area cultivated with soybean has been cultivated with GR varieties since 2013 [5]. Approximately 36% of this crop is planted with first-generation (RR) varieties and 60% is planted with second-generation (RR2) varieties [6], which also exhibit resistance to Lepidoptera. In both GM technologies, glyphosate resistance is achieved through the insertion of the cp4-epsps gene from Agrobacterium sp. This gene encodes an isoform of 5-enolpyruvylshikimate-3-phosphate synthetase (EPSPS), which is unaffected by glyphosate [7]. RR cultivars carry the original cp4-epsps gene driven by the 35S promoter, whereas RR2 cultivars harbor a modified cp4-epsps gene controlled by the FMV (fig mosaic virus) promoter [8].

Low glyphosate doses (considering the recommended field doses for weed control) are recognized for stimulating plant growth, a phenomenon known as hormesis [9,10]. Several crops, including soybean, corn, sugar cane, coffee, eucalyptus, and pine, have shown increased biomass, plant height, and yield with low-dose glyphosate foliar applications [11,12,13,14]; therefore, hormesis can be a valuable tool for boosting crop yields [15,16]. However, certain weed species may also exhibit growth stimuli with low doses of glyphosate [17,18], promoting the development of herbicide resistance [19]. This poses a challenge for applying low glyphosate doses via foliar in the field to induce hormesis and improve crop yield. Therefore, testing different methods, such as seed treatment, is necessary to explore effective ways of applying low doses of glyphosate to soybeans.

Pesticide seed treatment is a common agricultural practice to safeguard seeds and enhance germination, seedling vigor, and overall crop health [20]. In 2013, seed industry reports indicated that approximately 70% of sold soybean seeds received a pesticide treatment [21]. Considering the potential to increase soybean yield through glyphosate hormesis without encroaching on natural areas and without promoting weed growth or resistance evolution, it is vital to conduct cost-effective studies at the field level. In this context, seed treatment with low glyphosate doses emerges as an alternative for assessing the field yield of RR and RR2 cultivars. It is well documented that the mineral composition of GR crops remains unchanged with recommended glyphosate doses for weed control [22]; however, evidence suggests that low doses inducing hormesis can affect foliar nutrition in common beans [11]. Therefore, evaluating this parameter in soybeans is important to discard or document any potential changes in their mineral composition.

Previous greenhouse experiments conducted by this research group demonstrated that the safe range of glyphosate doses inducing hormesis via seed treatment varies among soybean cultivars. For non-GM cultivars, the range was from 45 to 90 g ae ha⁻¹; for RR cultivars, it was from 90 to 180 g ae ha⁻¹; and for RR2 cultivars, it was from 90 to 360 g ae ha⁻¹ [23]. The aim of this study was to assess the growth, leaf nutritional composition, and yield of RR and RR2 cultivars. These GR cultivars were planted in various soybean-producing locations in Brazil and treated with a common low hormesis-inducing dose of glyphosate (90 g ae ha⁻¹), applied through foliar, seed, and seed + foliar methods.

2. Materials and Methods

2.1. Description of Study Sites

The experiments were conducted in the field at four locations: Assis Chateaubriand—PR (24°14′47.2″ S 53°37′43.7″ W), Botucatu—SP (22°49′33.8″ S 48°25′47.8″ W), Palotina—PR (24°11′43.6″ S 53°48′26.3″ W), and Marechal Cândido Rondon (MC Rondon)—PR (24°41′01.0″ S 54°07′11.8″ W) during the 2018/19 harvest season.

After soil analysis in each location (

Table 1. Chemical and physical characteristics of the soil in the experimental areas Assis Chateaubriand, Botucatu, Marechal Cândido Rondon (MC Rondon), and Palotina, Brazil, in the 0–20 cm layer.

Local	pH	OM	P	Al3+	K	Ca	Mg	SB	CTC	V%	Sand	Silt	Clay	Texture
				mmolc/dm3							g kg−1
Assis Chateaubriand	6.2	21.9	22.2	0	8.7	11.5	5.9	26	111	58	230	245	525	Clayish
Botucatu	5.4	17.0	20.4	0	6.1	18.0	10	35	101	60	195	278	527	Clayish
MC Rondon	5.6	36.4	22.3	0	4.9	12.6	5.1	23	116	58	156	235	606	Clayish
Palotina	5.1	18.2	20.8	0	5.1	10.5	2.5	18	94	55	310	201	489	Clayish

pH (CaCl₂); OM—organic matter (g/dm³); P_resine (mg/dm³).

), liming was conducted before soybean sowing to achieve 60% base saturation, followed by heavy harrowing. The planting dates were 25 September for Assis Chateaubriand, 15 October for Botucatu, 1 October for MC Rondon, and 17 December for Palotina, all in 2018. Fertilization, involving an NPK formulation of 08-20-16, was applied in all locations at the time of sowing (330 thousand plants ha⁻¹ or ~60 kg seed ha⁻¹). Liming and fertilization were performed in accordance with the recommendations obtained from the interpretation of the soil analysis [24].

The experiment was conducted in a randomized block design with four replications, using four GR soybean cultivars: two RR (BMX-Tornado and N5909) and two RR2 (M5917-IPRO and M5838-IPRO). Seeds were purchased from Brasmax Genética (Cambé, Brazil). The plots (experimental units) consisted of six soybean rows, spaced 0.45 cm apart and 5 m in length.

2.2. Glyphosate Hormetic Treatments

The dose of glyphosate found to induce hormesis with the potential to increase soybean yield was 90 g ae ha⁻¹, a safe dose for plants of both types of GR cultivar at the V3 stage [23]. The glyphosate formulation used was based on potassium salt (Roundup Transorb R, 480 g ae L⁻¹, Monsanto do Brasil Ltd.a., Sao Paulo, Brazil), applied via foliar, seed, or seed + foliar. A control group without herbicide application was included.

Foliar applications on soybean plants were carried out at the V3 phenological stage of plants (three fully expanded trifoliate leaves), using a backpack sprayer pressurized with CO₂ (constant pressure of 2 bar or 29 PSI) to deliver 200 L ha⁻¹ at a speed of 1 m s⁻¹. The sprayer had an application bar with six XR110.02 nozzles, spaced 50 cm apart, and spraying was conducted at a height of 50 cm from the target. The meteorological conditions at the time of application were 28, 25, 29, and 27 °C, with 60, 64, 66, and 56% relative humidity and 1.3, 2.6, 1.0, and 1.5 km h⁻¹ wind speed in MC Rondon, Palotina, Assis Chateaubriand, and Botucatu, respectively. Meteorological data on maximum and minimum temperature and precipitation during the experiments are shown in Figure S1.

For each experimental site, sufficient seeds of each soybean variety (approximately 2 kg) for all plots were treated with glyphosate by the ‘on-farm’ method. Glyphosate (90 g ae ha⁻¹–3.125 mL of commercial glyphosate formulation kg⁻¹ seed), an organosilicon adjuvant (Silwet^® at a concentration of 0.05%), a seed protector product (2 mL kg⁻¹ seed), and an inoculant (5 g kg⁻¹ seed) were added to a bag containing the seeds. The mixture was shaken vigorously until all seeds were uniformly treated. The seed protector product, Standak^® Top UBS (BASF S.A., Guaratinguetá, Brazil), based on pyraclostrobin + methyl thiophanate + fipronil, was used to prevent seed rot caused by fungi and insect predation in the initial stage, and the inoculant, Brasilec TS In-Box (Forquimica, Cambira, Brazil), containing the strains Semia 587 and Semia 5019 of Bradyrhizobium elkani (5 × 10⁹ UFC g⁻¹) was used to induce and improve root nodulation. For controls (0 g ae ha⁻¹ glyphosate), seeds were treated solely with the adjuvant, seed protector product, and inoculant, all combined in a single treatment. After treatment, the seeds were left in the shade to dry for 24 h and subsequently sown. In the case of the seed + foliar treatment, 90 g ae ha⁻¹ was applied in each application, totaling 180 g ae ha⁻¹ glyphosate for this treatment.

For weed control, glyphosate (720 g ae ha⁻¹) was applied using the equipment and conditions previously described at the V4 stage. This herbicide was chosen because it accompanies RR technology, but, to prevent this application from interfering with our results, protective hoods were placed on the application bar to avoid the direct contact of glyphosate with the soybean foliage.

2.3. Evaluated Parameters

The variables evaluated were final plant height (cm), aerial part dry mass (g), root dry mass (g), total dry mass (g), number of root nodules, dry mass of nodules (mg), macro and micronutrient content (g kg⁻¹), number of pods per plant, and yield (kg ha⁻¹). At the R1 stage, five plants per plot were collected to determine the dry mass of the aerial parts and roots and the total, as well as the number and dry mass of nodules. However, we could not evaluate these variables in Botucatu. The aerial parts were cut at the ground level and stored in paper bags. For roots and nodules, the soil around the plant was carefully removed with a stream of water to avoid damaging them, and they were also placed in paper bags. All samples were dried in a forced air oven at 60 °C until a constant weight was reached. Subsequently, the number of nodules was quantified and the samples were weighed.

The plant height was evaluated with a tape measure, measuring from the soil surface to the last fully expanded leaf in 10 randomly chosen plants in each plot. The number of pods was counted at stage R7 (physiological maturity) on 10 plants chosen at random per plot. Ten trifoliate leaves from the third apex node to the soil surface at R1 were collected to quantify the macro and micronutrient contents (N, P, K, Ca, Mg, S, B, Cu, Fe, Mn, and Zn), following the methodologies described by Malavolta et al. [25]. Yield was determined by collecting and threshing the two 4 m long central rows of the plots with an experimental thresher. The samples were cleaned with sieves and placed in paper bags. Based on the weight of the seeds, the yield was estimated for each treatment and replication, correcting the degree of seed moisture to 13%.

2.4. Statistical Analysis

For each variable, a variance stability test was carried out to consolidate the data across treatments, irrespective of location. However, due to significant variations, the data underwent a 3-way ANOVA to assess the impact of each factor (location, cultivar, and application method) on the analyzed parameter, considering substantial variations. Normal distribution of errors and homogeneous variance assumptions were visually verified for all tests. Statistical significance was set at p < 0.05 and mean comparisons were conducted using Tukey’s HSD test, with a confidence level of 95%. Furthermore, the confidence interval (CI) was calculated for all variables analyzed using the following equation: CI = (t × SD)/√n, where t = tabulated t-value at 5%, SD = standard deviation, and √n = square root of the number of repetitions. The analyses were conducted in Statistix 9 (Analitycal Software, Tallahassee, FL, USA), and data were plotted using Sigma Plot 12.5.

3. Results

3.1. Dry Mass

The total dry weight per plant of GR soybean plants in the R1 stage ranged from 4.0 to 8.0 g, with differences between locations, cultivars, and treatments. The average total dry mass recorded in Assis Chateaubriand (7.1 g) and MC Rondon (6.7 g) was similar and substantially higher than in Polotina (4.7 g). While the differences between GR cultivars were small at each location, the RR cultivar N5909 exhibited the lowest dry mass in Assis Chateaubriand (6.6 g) and Polotina (4.4 g), whereas the RR2 cultivar BMX-Tornado recorded the lowest dry mass in MC Rondon (6.3 g). In most cases, glyphosate treatments stimulated biomass production compared to their respective controls. The foliar application displayed irregular and modest increases, ranging from 5 to 11%, while the seed and seed + foliar treatments showed consistent increases, ranging between 7 and 21%. For instance, in Assis Chateaubriand, the RR2 cultivar M5917-IPRO under the seed + foliar treatment weighed 8.0 g, while the control weighed 6.6 g (Figure 1).

Evaluating the plant sections separately, the aerial parts showed increases ranging from 3.1 to 5.7 g, aligning with the trend observed for the total mass. However, the dry mass of roots (ranging from 0.9 to 2.3 g) varied at each experimental site. The average root weight recorded in Assis Chateaubriand, MC Rondon, and Polotina was 2.1, 1.5, and 1.0 g, respectively. When glyphosate contributed to the increase in root mass, it ranged from 5 to 22% for the foliar treatment and from 7 to 33% for the seed and seed + foliar treatments. However, the RR cultivar BMX-Tornado presented a remarkable increase in root mass, ranging from 50 to 60% in MC Rondon across all treatments (Figure S2).

3.2. Root Nodules

The number of nodules of GR plants ranged from 21 to 36 nodules plant⁻¹, with no differences observed between experimental sites and among cultivars within each site, except in Palotina, where the RR2 cultivar M5917-IPRO displayed the highest number of nodules per plant (29.7 nodules plant⁻¹). Glyphosate was associated with these variations since the three treatments stimulated the formation of nodules by up to 36% compared to their respective controls. Only the RR2 cultivar M5838-IPRO in MC Rondon and the cultivars N5905 and BMX-Tornado in Polotina presented a number of nodules (26 plant⁻¹) in the foliar treatment similar to their respective controls (Figure 2). In contrast, the dry mass of nodules (ranging from 170 to 367 mg plant⁻¹) differed among locations, cultivars, and treatments. The average nodule dry mass recorded in Assis Chateaubriand, MC Rondon, and Palotina was 279, 240, and 198 mg, respectively. Within each location, the highest weight was recorded in the RR2 cultivar M5917-IPRO, followed by the RR cultivar BMX-Tornado. Although, in most cases, glyphosate treatments increased the nodule dry mass, the treatment that caused the greatest increase was seed + foliar (up to 56%), followed by seed (up to 41%) (Figure 2).

3.3. Nutrient Composition

The nutritional composition of the leaves of GR soybean plants varied among locations. In certain instances, differences were observed between cultivars, while glyphosate treatments rarely altered the macro and micronutrient contents.

Variations were notable for macronutrients N and P and micronutrients Br and Fe between locations. The lowest N concentration (42 g kg⁻¹) and the highest P and Br concentrations (4.8 g kg⁻¹ and 58 g kg⁻¹, respectively) were registered in Botucatu. In Polotina, there were no differences in N content between cultivars or glyphosate treatments; in Assis Chateaubriand, there were differences among cultivars, with RR cultivars having the highest N content (48 g kg⁻¹). In Botucatu and MC Rondon, differences existed between cultivars and treatments. The RR2 cultivar M5917-IPRO had the highest N content (50 g kg⁻¹) in these areas, followed by the RR cultivar BMX-Tornado. The P content ranged from 3.5 to 6.0 g kg⁻¹. In Polotina, there were only differences between cultivars, with the RR cultivar BMX-Tornado showing the highest P content (4.7 g kg⁻¹). In the majority of locations, variations in P content were evident between cultivars and, in certain instances, between treatments. Specifically, in Botucatu, the cultivars RR2 and N5909 (RR) exhibited the highest P contents (ranging from 5.0 to 5.3 g kg⁻¹); in MC Rondon, only the RR2 cultivar M5917-IPRO recorded the highest P content (4.4 g kg⁻¹), while in Polotina, the BMX-Tornado stood out, with a P content of 4.5 g kg⁻¹ (

Table 2. Macro and micronutrient content (g kg⁻¹ dry mass) with large statistical differences for soybean RR2 (M5917-IPRO and M5838-IPRO) and RR (BMX-Tornado and N5909) cultivars treated with glyphosate via foliar (Fol.: 90 g ae ha⁻¹), seed (90 g ae ha⁻¹), and seed + foliar (S + F: 180 g ae ha⁻¹) in Assis Chateaubriand, Marechal Cândido Rondon, and Palotina, Brazil, in comparison to untreated control plants (Con.).

Nutrient	Cultivar	Assis Chateaubriand					Botucatu					Marechal Cândido Rondon					Polotina
		Con.	Fol.	Seed	S + F	$\overline{\mathbf{X}}$	Con.	Fol.	Seed	S + F	$\overline{\mathbf{X}}$	Con.	Fol.	Seed	S + F	$\overline{\mathbf{X}}$	Con.	Fol.	Seed	S + F	$\overline{\mathbf{X}}$
N		46 A					42 B					45 A					45 A
	M5917-IPRO	42	45	45	42	43 B	40 b	42 b	47 a	49 a	45 A	44 b	44 b	48 ab	50 b	47 A	46	46	46	46	46
	M5838-IPRO	41	43	46	43	43 B	39 a	33 b	45 a	48 a	41 BC	44 a	41 b	46 a	41 b	43 B	43	43	47	45	44
	N5909	46	46	52	49	48 A	38	41	38	40	39 C	46	45	46	43	45 AB	47	46	46	45	46
	BMX-Tornado	47	46	50	48	48 A	39 b	40 b	43 ab	48 a	43 AB	47	47	48	46	47 A	45	46	43	46	45
		4.3 B					4.8 A					4.1 B					4.3 B
P	M5917-IPRO	4.1	4.5	4.2	4.1	4.2 B	4.9 b	5.0 b	5.2 b	6.0 a	5.3 A	4.1 b	4.1 b	4.8 a	4.6 a	4.4 A	3.8 b	4.1 ab	4.3 ab	4.8 a	4.2 B
	M5838-IPRO	4.3	4.2	4.4	4.1	4.2 B	5.1	5.1	5.2	5.6	5.3 A	4.0	4.2	3.8	4.0	4.0 B	4.2	4.3	4.1	4.1	4.2 B
	N5909	4.1	4.0	4.0	3.9	4.0 B	4.4 b	4.8 ab	5.0 ab	5.7 a	5.0 A	3.6 b	3.7 b	4.0 ab	4.2 a	3.9 B	4.0	4.3	4.2	4.3	4.2 B
	BMX-Tornado	4.6	4.6	4.8	4.9	4.7 A	3.3 b	3.2 b	3.9 ab	4.5 a	3.7 B	3.5 b	3.6 b	4.6 a	4.7 a	4.1 B	4.1 b	4.2 b	4.8 a	5.0 a	4.5 A
Br		39 C					58 A					36 D					42 B
	M5917-IPRO	45	45	47	42	45 A	60 a	69 ab	76 a	84 a	72 A	38 b	44 ab	49 a	50 a	45 A	37 b	41 ab	46 a	47 a	43 B
	M5838-IPRO	40	48	43	43	44 A	55 ab	47	48	50	50 B	37	36	35	36	36 B	46 b	49 b	48 b	56 a	50 A
	N5909	33 b	32 b	38 ab	41 a	36 B	52 b	51	47	59	52 B	32	28	30	27	29 C	36	38	35	37	36 C
	BMX-Tornado	31	31	33	31	31 B	51 b	55 b	58 ab	61 a	56 B	26 b	29 b	35 a	39 a	32 BC	35 b	40 ab	42 a	45 a	41 B
Fe		173 C					256 A					257 A					205 B
	M5917-IPRO	188 b	175 b	251 a	253 a	217 A	244	267	250	254	254	239 b	234 b	273 a	270 a	254 B	214	221	198	213	212
	M5838-IPRO	169	196	166	157	172 B	254	263	253	279	262	249 b	292 ab	323 a	324 a	297 A	207	197	226	210	210
	N5909	149 b	144 b	183 a	150 b	157 BC	255	257	292	260	266	224 b	219 b	248 a	211 b	226 C	205	220	199	205	207
	BMX-Tornado	132 b	142 a	150 a	158 a	146 C	254	240	225	252	243	245	246	255	253	250 B	180	174	198	209	190

Mean ± confidence interval (5% probability). Same capital letters between locations, italic capital letters between cultivars, and lowercase letters between treatments within a cultivar did not differ from each other using the Tukey test (p ≤ 0.05).

).

The Br content exhibited substantial variation, reaching its peak at 58 g kg⁻¹ in Botucatu. In contrast, the averages for other locations ranged from 36 to 42 g kg⁻¹. Differences were notable among cultivars, with one or both RR2 cultivars having the highest Br contents. The Fe content ranged widely (132 to 324 g kg⁻¹), with Botucatu and MC Rondon presenting the highest concentrations (256.5 g kg⁻¹), while the lowest content was observed in Assis Chateaubriand (173 g kg⁻¹). In Botucatu and Polotina (205 g kg⁻¹), no differences were observed between cultivars or treatments. In Assis Chateaubriand and MC Rondon, the RR2 cultivars M5917-IPRO and M5838-IPRO, respectively, displayed the highest Fe contents. When differences between treatments were recorded, the highest concentrations of all these macro and micronutrients were observed in the treatments via seed or seed + foliar (Table S1).

3.4. Plant Height

The height of GR soybean plants varied among experimental sites and cultivars, with fewer variations among treatments. The tallest plants were observed in Assis Chateaubriand (82.5 cm), while the shortest were in Polotina (60.2 cm). In Polotina, no differences were found between cultivars, but the cultivar M5917-IPRO showed variations among treatments, recording the highest height in the control and seed + foliar treatments. In other locations, the tallest plants, reaching up to 81.3 cm, belonged to the RR cultivar BMX-Tornado, followed by plants of the RR2 cultivar M5917-IPRO (65.1–73.3 cm). In Botucatu, no differences were observed between treatments within each cultivar. When differences were present, plant height tended to decrease compared to the respective controls. In Assis Chateaubriand, the cultivar M5838-IPRO exhibited a height 8% lower than the control in the seed + foliar treatment, while, in MC Rondon, the M5917-IPRO and BMX-Tornado cultivars (RR2 and RR, respectively) showed reduced heights by 8.3 and 5.8%, respectively, in the foliar and seed + foliar treatments (Figure 3).

3.5. Pods and Yield

The pod number of GR soybean plants ranged from 36.7 to 72.1 pods plant⁻¹, depending on the experimental site, cultivar, and glyphosate treatment. The highest pod numbers were recorded in Assis Chateaubriand and Botucatu (61.5 and 63.8 pods plant⁻¹, respectively), while the lowest was in Polotina (45.1 pods plant⁻¹). In Polotina, the RR2 cultivar M5838-IPRO recorded the lowest pod number (39.9 pods plant⁻¹). In other locations, the other RR2 cultivar M5917-IPRO generally produced the fewest pods (ranging from 49 to 59.5 pods plant⁻¹), while the RR cultivar BMX-Tornado produced the most (61.1 to 67.5 pods plant⁻¹). Differences between treatments, when they occurred, increased pod numbers. In the foliar treatment, the pod number was irregular, ranging from −2 to +23%, sometimes similar or lower than the control, and at other times similar to the other treatments. For seed and seed + foliar treatments, the increases were consistent, ranging from 10 to 35% and 11 to 42%, respectively (Figure 4).

Soybean yield ranged from 1345 to 4691 kg ha⁻¹, varying mainly among locations. In MC Rondon and Polotina, all cultivars yielded below 2000 kg ha⁻¹, except BMX-Tornado in MC Rondon, which yielded 2718 kg ha⁻¹. In Assis Chateaubriand, the yield ranged from 2283 to 2476 kg ha⁻¹, with N5909 and BMX-Tornado presenting the lowest and highest yields, respectively. In Botucatu, the yield was 4143 kg ha⁻¹, with BMX-Tornado also being the most productive cultivar (4547 kg ha⁻¹). When treatments influenced yield, there was an increase compared to the controls. In the foliar treatment, the increases were low and irregular, ranging from 5 to 20%. In the seed and seed + foliar treatments, the increases were consistent, ranging from 7 to 37% and 13 to 38%, respectively (Figure 5).

4. Discussion

The agronomic characteristics and yield of GR soybeans showed variability across different locations, with further variations observed within each location based on the cultivar and/or glyphosate treatment. In the locations of Paraná, there were periods of low water precipitation and high temperatures during the experiments, leading to water stress and substantially reduced soybean yield compared to Botucatu in São Paulo State, where there was no water restriction (Figure S1). Environmental conditions interfere with the response of plants to low doses of glyphosate [11,26]. Despite these reductions, glyphosate treatments increased soybean yield in most cases. In common beans, low doses of glyphosate altered physiology and increased yield depending on planting time, with an increase of up to 23% in the winter season and up to 109% in the rainy season [27]. This increase is linked to the hormetic effects of glyphosate, which mitigates the effects of water stress by activating antioxidant enzymes [28]. Favorable environments support physiological responses to the hormetic effect of glyphosate by, for example, attempting to compensate for the minor chemical stress induced in plants [29,30], reducing lignin levels, increasing growth [31], and enhancing P absorption [32,33].

Among soybean cultivars, the RR2 cultivar M5917-IPRO generally displayed superior agronomic characteristics across most locations in the absence of glyphosate treatment. On the other hand, the RR cultivar BMX-Tornado (RR2) exhibited the greatest response to hormetic stimuli induced by the low dose of this herbicide. This response manifested in increased nodules per plant, height, pod number, and overall yield. For M5838-IPRO (RR) and N5909 (RR2) cultivars, glyphosate led to an increase in certain variables, but the most critical factors (pod number and yield) were location-dependent.

When glyphosate treatments stimulated the growth and increased the yield of GR soybean, these effects were low and irregular for foliar treatment, while they were high and consistent for seed or seed + foliar treatments. A foliar treatment relies heavily on environmental conditions to trigger hormesis. During foliar applications, underdoses and overdoses of the expected concentration occur [31]. This means that some soybean plants received concentrations both above and below 90 g ae ha⁻¹, resulting in irregular hormetic responses. In contrast, seed treatments provide greater uniformity, cost-effectiveness, and reduced sensitivity to environmental conditions [34], having a milder impact on hormetic responses since the herbicide contacts the seedlings during germination and emergence. The better results in mass and the number of nodules with the seed + foliar treatment may be attributed to the predisposition of the plants to glyphosate induced by the seed treatment. This predisposition enhanced hormetic responses to the subsequent foliar application [35]. Similarly, lettuce plants exhibited increased root growth in response to low doses of glyphosate when preconditioned to low methanol stress [14]. However, the seed + foliar treatment with low doses of glyphosate that induced hormesis is only applicable in weed-free areas to prevent the foliar application from contributing to the selection of glyphosate-resistant weed biotypes.

Biomass accumulation data revealed that glyphosate increased biomass in both the aerial parts and roots of GR soybean plants. This indicates that resource allocation occurs throughout the plant in response to the hormetic stimuli of glyphosate [12]. Seven days after application, low glyphosate doses increased both root and aerial biomass in newly germinated soybean seedlings [36]. This biomass increase may be attributed to a reduction in lignin levels following glyphosate application, redirecting photoassimilates from lignin production to other parts of the plant [10]. A greater root presence enhances nutrient and water absorption in soybeans, optimizing soil exploitation [37,38]. Foliar and seed treatments increased the root biomass and nodulation of GR soybean plants by up to 36% in most locations compared to controls, underscoring the significant impact of the herbicide on nodule development. Low glyphosate doses were found to promote mycorrhizal colonization due to enhanced isoflavonoid production by soybeans [39]. Applying 90 g ha⁻¹ of glyphosate during flowering increased both the number and mass of soybean nodules [39]. A dose of 105 g ha⁻¹ applied 70 days after sowing increased nodule mass in glyphosate-susceptible soybeans compared to the untreated control [40]. The greater nodulation was attributed to increased root growth in soybeans treated with low glyphosate doses [41].

The increased root mass and nodule numbers resulting from low-dose glyphosate treatments contributed to improved soybean yield in water-stressed areas in Paraná. Low doses of glyphosate (7.2 g to 36.0 g ae ha⁻¹) changed the leaf nutrient content in common beans cultivated in Botucatu during the dry season. However, there were no differences in nutrient composition during the rainy season [11]. In the case of soybean, the glyphosate dose of 105 g ae ha⁻¹ reduced the Fe and N contents in leaves and grains in a glyphosate-susceptible cultivar [40,42], while doses from 0.8 to 8.6 g ae ha⁻¹ increased P content and decreased Ca, Mg, Mn, and Fe contents in leaves [32]. In this study, the content of macro and micronutrients in the leaves of GR soybean plants was directly related to the experimental locations. Although the soil analysis was not robust, the concentrations of K, Ca, and Mg were highest in Botucatu and the lowest in Polotina. This accounted for the variations in the soybean yield observed in these locations. For nutrients such as N, P, Br, and Fe, which exhibited changes in their concentrations in leaves associated with glyphosate treatments in some instances, the differences generally showed an increase.

The nutrient most affected by glyphosate application in GR soybean cultivars was P. The elevation in P contents following glyphosate application was associated with the competition between the phosphate group of glyphosate and P for transport into the cell through phosphate transporters [43]. This competition signaled a lack of P in the cells, triggering the expression of highly specific transporter genes, which are more efficient in transporting P [42], favoring the increased absorption and allocation of this nutrient [33,44]. Consequently, soybean plants treated with glyphosate exhibited a higher concentration of P. Studies with low doses of glyphosate have also demonstrated increased P absorption in soybean and eucalyptus plants compared to untreated plants [45]. However, in a test with P deficiency, no positive glyphosate hormesis responses were found when compared to a control with a high P content [46]. This suggests that P dynamics in plants may be one of several processes involved in glyphosate hormesis, and the presence of this nutrient in the soil or substrate may determine the expression of hormetic effects.

The number of pods significantly impacts soybean yield [47]. Applying glyphosate increased pod numbers and yield across different locations and GR soybean cultivars. However, the yield increase was not solely due to more pods but resulted from a complex interaction of factors triggered by the stimulating effect of low-dose glyphosate. This included greater biomass, increased nodules, and, in some cases, higher concentrations of macro and micronutrients. This emphasizes the complexity of factors contributing to the glyphosate hormetic response in plants [10,11,15,16]. In the literature, there is no evidence that glyphosate hormesis improves soybean yield in the field. In sugarcane, low doses of glyphosate increased the P content in leaves and enhanced growth and yield [14].

5. Conclusions

Glyphosate hormesis treatments influenced the growth and agronomic characteristics of soybeans, but differences primarily depended on locations, cultivars, and treatment methods. Seed and seed + foliar treatments of glyphosate induced hormetic responses in GR soybean cultivars in most cases and regardless of location, leading to increased biomass accumulation, improved resource allocation, enhanced root development and nodulation, increased plant height, and higher yield. Additionally, glyphosate treatments did not alter the nutritional composition of soybeans. Any observed differences primarily stemmed from location-specific characteristics. The eventual superior results observed in seed + foliar treatments may be attributed to seed-induced predisposition. This study highlights the complex interplay of factors influencing the glyphosate hormetic response in soybeans and introduces alternative and viable methods for utilizing herbicide hormesis as a tool to increase crop yield at the field level.