English (pdf)
Article in xml format
Article references
Automatic translation
Send this article by e-mail
Cited by SciELO 
Similars in
SciELO 
Permalink
Introduction
Glyphosate is a non-selective broad-spectrum herbicide, widely used on global scale 3,12. The extended use of glyphosate has been attributed to its efficacy on control weed species 26, but an increasing number of reports suggest negative side effects on non-target plants 5. Many studies have found detrimental effects of glyphosate on seed germination and seedling growth of several crop species 1,4,8,28,31. However, the results are not conclusive because other studies have found little or negligible effects on crop species 7,18,27.
In cattle or sheep grazing systems, glyphosate is sprayed to suppress the existing vegetation before seeding or reseeding pastures or forage crops by no-tillage methods as direct drilling or broadcasting seeds 2,34. However, little information is available about the effect of glyphosate on seed germination and seedling growth of forage species because most studies about the exposure of glyphosate on the root-zone were carried out on crop species and rarely in a real soil substrate.
Numerous studies consider that glyphosate toxicity to non-target plants in soils is marginal because it is inactivated by microbial degradation and adsorption by soil constituents 3. Glyphosate is a polar compound which adsorbs strongly to soil iron and aluminum oxides and clays 24. Therefore, the adsorption of glyphosate and aminomethylphosphonic acid (AMPA) increased strongly with iron and aluminum content of soils and decreased with increasing soil organic carbon content. This would indicate that glyphosate and AMPA are mainly adsorbed by clay minerals, while soil organic matter competes for adsorption sites and inhibits adsorption 14. Such soil binding is reversible, and after desorption glyphosate is degraded by various bacteria to AMPA, its main secondary metabolite, and ultimately to inorganic phosphate, ammonia and carbon dioxide 3,4. Therefore, the labile pool of glyphosate in soils to be absorbed by plant roots could be dramatically reduced and, glyphosate toxicity over non-target plants would be negligible 15. Supported by these evidences, it was globally widespread that glyphosate is inactive in the soil, and therefore its phytotoxic effect could be considered as negligible. However, other studies detected residues of glyphosate and AMPA in soil solution after six months after spraying 32. A similar degradation process of glyphosate to AMPA also occurred in some plant species 29, a portion of these residues in soil would come from the glyphosate absorbed by plants, translocated towards the roots and underground organs and then exuded to the surrounding solution 21. Therefore, an important source of soil glyphosate or AMPA is root exudation from sprayed plants and its release from dead plants, which implies a toxicity risk to non-target plants due to rhizosphere transfer 16.
Due to their chemical similarities, both glyphosate and AMPA can be taken up by roots from the soil solution 16, so affecting the growth of seedlings. In forage species, only one report stated that the incorporation of glyphosate before sowing into a Vertic Argiudol reduced chlorophyll content and above and belowground growth in seedlings of the legume Lotus corniculatus8. In accordance, there is evidence that rootlets of soybean and maize seeds that germinate after spraying could absorb secondary metabolites of glyphosate, mainly AMPA exuded by roots of sprayed plants 31. Besides, glyphosate degradation resulted two to six times slower when it was previously absorbed by plants than when it was directly applied to the soil 10. The elapsed-time between glyphosate spraying and sowing was directly related to sunflower seedling growth and biomass production, and glyphosate had yet detrimental effects at the largest elapsed-time of 21 days 33. Glyphosate markedly inhibited growth of rootlets and shoots of Raphanus sativus seedlings exposed to the herbicide in the root-zone after 7 and 14-days of exposure 18. Negative effect on seed germination was found when seeds of legume species were in direct contact with glyphosate in water solution 17,23. In summary, contrary to the idea that glyphosate is inactivated in soil, several studies suggest that glyphosate spraying on the pre-existing vegetation may cause deleterious effects on seed germination, seedling emergence and growth of the subsequent crop.
Therefore, the aim of this study was to evaluate the effect of glyphosate spraying on germination, seedling emergence and seedling growth of four greatly extended forage species. We predicted that negative effects of glyphosate on seed germination and establishment would be greater when glyphosate is sprayed on pre-existing vegetation than on bare soil, and that the elapsed-time between spraying and seeding will be positively related to germination, seedling emergence and growth of forage seedlings.
Materials and methods
Experimental design
We extracted soil from paddocks of a livestock farm located in the center of the Flooding Pampa region (36°40’12” S, 59°32’10” W, 80 m a. s. l.). We selected paddocks covered by native grasslands that had never been sprayed with any herbicide. The dominant soil belongs to the series General Guido, typic Natraquoll, characterized by an acidic (pH 6.8), non-saline A1 horizon 12 cm-depth and a clayey and sodic B2t horizon (pH 8.2). Soil was extracted from the top horizon, sieved to remove vegetation and seeds, and then distributed in 400 plastic containers of 13 x 17 x 25 cm. Containers were maintained in a greenhouse with controlled temperature (20-25°C), natural light and regular watering throughout the experimental period. In April 2010, 20 seeds of each forage species L. tenuis, T. repens, P. dilatatum and F. arundinacea were sown in 100 containers per species and covering them with a thin layer of the same soil.
To compare the effects of glyphosate sprayed on pre-existing vegetation cover (C+) or on bare soil (C-) on germination, seedling emergence and seedling growth of the four forage species, we generated a vegetation cover of Lolium multiflorum Lam plants. For this, we harvested seeds of L. multiflorum in December 2009 from the same grassland paddocks where soil was extracted. Twenty-five seeds of L. multiflorum per container were sown in 200 containers 45 days before glyphosate spraying. Plants of L. multiflorum were thinned to 25 plants per container and periodically cut to maintain a similar height of 12 cm. In the 200 containers under C+ level, L. multiflorum plants were cut to ground level two days after sowing to avoid confounding effects of competition between L. multiflorum plants and forage seedlings. Other 200 containers were not sown with L. multiflorum to obtain the C- level.
To evaluate whether the elapsed-time between glyphosate spraying and sowing was positively related to germination, seedling emergence and growth of seedlings, glyphosate was sprayed 1 (G1), 30 (G30), 60 (G60) and 90 (G90) days before sowing the forage seeds or was not sprayed (G0 - control treatment sprayed with water 1day before sowing) 9.
A monoammonium salt of glyphosate (1440 g ae ha-1) was applied at a dose of 2.88 L of glyphosate ha-1 with a CO2 pressurized backpack sprayer fitted with a TT11001 nozzle delivering 100 L ha-1 at 200 kPa 36. The levels G1, G30, G60 and G90 consisted in 40 containers that were kept without vegetation cover (C-) and 40 containers that were sown with 25 seeds of L. multiflorum (C+) 60 days before glyphosate spraying. Control level G0 consisted in 80 containers (40 C+ and 40 C-) sprayed with the same amount of water without glyphosate one day before the sowing of forage seeds.
In summary, a completely randomized factorial design with 5 replications was performed. The two main factors were 1) Cover on which glyphosate was sprayed (C), at 2 levels: presence (C+) or absence (C-) of vegetation cover; and 2) Spraying of glyphosate before sowing (G), at 5 levels: sprayed 1(G1), 30 (G30), 60 (G60) and 90 (G90) days before sowing or not sprayed (G0). The same experiment was performed for each forage species and we analyzed the responses of each forage species to glyphosate treatments but did not compare among species due the great morphological and physiological differences among them.
Plant material
We selected two forage legume species, Lotus tenuis Waldst. cv. Aguapé and Kit. Ex Willd and Trifolium repens L. cv. El Lucero and two forage grass species Paspalum dilatatum Poir. (cv. Relincho) and Festuca arundinacea (Schreb.). These species are worldwide used in mixed pastures and are also naturalized in native grasslands of temperate regions 13. T. repens and F. arundinacea grow during the cool-season period, while L. tenuis and P. dilatatum grow during the warm-season period. Seeds of L. tenuis, T. repens, P. dilatatum and F. arundinacea were provided by commercial seed breeders.
Response variables
The number of emerged seedlings was daily recorded until twenty consecutive days without new emergencies. This period averaged 40 days after sowing. After 60 days from sowing, all seedlings were carefully harvested and seeds remaining in the soil were exhumed by sieving the soil through two sieves of 2.0 and 0.2 mm and examined under an optical microscope to determine whether germination had initiated (i.e. rootlet appearance) 30. The exhumed seeds that did not germinate (i.e. without visible rootlet) were evaluated by the tetrazolium test to determine their viability 19. The response variables related to seed germination and seedling emergence were expressed in percentage and calculated as following: a) Seed germination (%) as the number of established seedlings plus the number of exhumed germinated seeds in relation to the number of seeds sown of each species, b) Viable seeds (%) as the number of established seedlings plus the number of exhumed germinated seeds plus the number of viable not germinated seeds in relation to the number of seeds sown of each species, and Seedling emergence (%) as the number of established seedlings in relation to the number of seeds sown of each species 30.
Seedling morphological traits were assessed after seedling removal by measuring the number of leaves of the unique tiller and the length of the third leaf (lamina plus sheath) in the grasses P. dilatatum and F. arundinacea, the number and average length of the stems in L. tenuis and the number of leaves and the average length of the petioles in T. repens. The length of the seminal root of the four species was also measured. All length measurements were carried out by using a mechanical caliber. Aboveground and belowground biomass of seedlings were determined by oven drying until constant weight.
Statistical analyses
Variables were analyzed separately for each species through a factorial two-way ANOVAs at α=0.05, being the main factors C and G. Tukey tests were applied to post-hoc comparison among means. Variables that involve percentages (seed germination, seedling emergence and viable seeds) were arcsine square-root transformed before analysis to meet ANOVA assumptions. Normality and homogeneity of variances were previously checked performing Kolmogorov-Smirnov and Levene and Brown-Forsythe tests. The results that involve percentages are shown as non-transformed mean + 1 SD. Statistical analyses were performed using the package STATISTICA for Windows (StatSoft,Tulsa, OK, USA).
Results
Effects of glyphosate spraying on seed germination and viability
Germination and seed viability were not affected by the main factors C and G or their interaction. Germination was 50.6 +12.3% for L. tenuis, 57.3+5.9 % for T. repens, 77.3+13.0% for P. dilatatum and 77.1+11.5% for F. arundinacea, while seed viability was 84.96+7.37% for L. tenuis, 82.16+6.28% for T. repens, 87.69+6.81% for P. dilatatum and 90.17+3.25% for F. arundinacea.
Effects of glyphosate spraying on seedling emergence
Seedling emergence of the four forage species was clearly affected by the interaction between them. Seedling emergence showed a consistent pattern among species: it was lower under C+ at G1, G30 and G60, than under C+ at G90 and G0 or under C- at any G level (Figure 1, page 40).
Vertical bars indicate 1 SD of the means. Different letters indicate significant differences provided by the Tukey test of the interaction factor (P< 0.05). Insert results of ANOVA test: Main factors are Cover on which glyphosate was sprayed (C, at two levels C+ and C-) and Spraying of glyphosate before sowing (G, at five levels G1, G30, G60, G90 or G0). Significant levels (SL) are indicated with an asterisk (*) when P<0.05 and with two asterisks (**) when P<0.01. Las barras verticales indican 1 desvío estándar de las medias. Letras distintas indican diferencias significativas según el test de Tukey (P<0,05) para el factor interacción. Se insertaron los resultados del ANOVA: Los factores principales son Cobertura sobre la que se pulverizó glifosato (C, a dos niveles: C+y C-) y Pulverización de glifosato antes de la siembra (G, a cinco niveles: G1, G30, G60, G90 or G0). El nivel de significancia (SL) se indica con un asterisco (*) cuando P<0,05 y con dos asteriscos (**) cuando P<0,01.
Figura 1: Emergencia de plántulas (%) de Lotus tenuis (a), Trifolium repens (b), Paspalum dilatatum (c) y Festuca arundinacea (d) pulverizadas con glifosato sobre la vegetación pre-existente de L. multiflorum (C+; barras naranjas) o sobre suelo desnudo (C-; barras blancas) 1, 30, 60, 90 días antes de la siembra y el tratamiento control no pulverizado (G1, G30, G60, G90 y G0).
Therefore, glyphosate sprayed on preexisting vegetation cover 1 to 60 days before sowing caused a negative effect on seedling emergence of different magnitude depending on the species. Seedling emergence of L. tenuis under C+ at G1, G30 and G60 was 45% lower than under C+ at G90 and G0 and C- at G1, G30, G60, G90 and G0 (26.1+10% vs. 47.4+11.5% respectively Figure 1 a, page 40). T. repens showed the lowest seedling emergence under C+ at G1 and G30 (21.1+6.9%), which was almost 63% lower than under C+ at G90 and G0 and C- at any G levels (56.5+3.7%) (Figure 1 b, page 40). Seedling emergence of P. dilatatum under C+ at G1, G30 and G60 was 35% lower than under C+ at G90 and G0 and C- at any G levels (49.9+13.9%vs 76.6+14.8% respectively, Figure 1 c, page 40). Seedling emergence of F. arundinacea under C+ at G1, G30 and G60 was almost 40% lower than under C+ at G90 and G0 and C- at any G levels (45.9 +14.5% vs. 75.7+12.1% respectively Figure 1 d, page 40).
Effects of glyphosate on seedling biomass
Above and belowground biomass of legumes seedlings was affected by the main factors C and G or by the interaction between them. Under C+ at G1 and G30, aboveground biomass of L. tenuis was 83% lower and above and belowground biomass of T. repens seedlings was 51% lower than under C+ at G90 and G0 and under C- at any G level (0.0169 +0,0069 g. vs. 0.0987+0.0393 g; 0.0105+0.0047 vs. 0.0213 +0.0054 g and 0.0067+0,0037 vs. 0.0136+0.0042 g respectively, Figure 2 a and b, page 41).
Vertical bars indicate 1 SD of the means. Different letters indicate significant differences provided by the Tukey test of the interaction factor (P< 0.05). Insert results of ANOVA test: Main factors are Cover on which glyphosate was sprayed (C, at two levels C+ and C-) and Spraying of glyphosate before sowing (G, at five levels G1, G30, G60, G90 or G0). Significant levels (SL) are indicated with an asterisk (*) when P<0.05 and with two asterisks (**) when P<0.01. Las barras verticales indican 1 desvío estándar de las medias. Letras distintas indican diferencias significativas según el test de Tukey (P< 0,05) para el factor interacción. Se insertaron los resultados del ANOVA: Los factores principales son Cobertura sobre la que se pulverizó glifosato (C, a dos niveles: C+y C-) y Pulverización de glifosato antes de la siembra (G, a cinco niveles: G1, G30, G60, G90 or G0). El nivel de significancia (SL) se indica con un asterisco (*) cuando P<0,05 y con dos asteriscos (**) cuando P<0,01.
Figura 2: Biomasa aérea (barras blancas) y subterránea (barras naranjas) de las plántulas de Lotus tenuis (a), Trifolium repens (b), Paspalum dilatatum (c) y Festuca arundinacea (d) pulverizadas con glifosato sobre la vegetación pre-existente de L. multiflorum (C+; barras naranjas) o sobre suelo desnudo (C-; barras blancas) 1, 30, 60, 90 días antes de la siembra y el tratamiento control no pulverizado (G1, G30, G60, G90 y G0).
Belowground biomass of L. tenuis seedlings was lower under C+ than under C- and at G1 and G30 respect to G90 and G0 (Figure 2 a, page 41). Under C+ at G60, aboveground biomass of L. tenuis seedlings was higher than under C+ at G1 and G30 but lower than under C+ at G90 and G0 and under C- at any G level (Figure 2 a, page 41), while above and belowground biomass of T. repens seedlings did not differ respect to the other G levels (Figure 2 b, page 41). Belowground -but not aboveground- biomass seedling of grasses was also affected by the main factors C and G and by the interaction between them. Belowground biomass of P. dilatatum seedlings was 46% lower under C+ at G1 and G30 than under C+ at G90 and G0 and under C- at any G level (0.0122+0.0050 g vs. 0.0224 + 0.0049 g, (Figure 2 c, page 41). Belowground biomass of F. arundinacea was 40% lower under C+ at G1 and G30 than under C+ at G60, G90 and G0 and C- at any G level (0.0148 +0.0033 g vs. 0.0248+0.0047 g, (Figure 2 d, page 41).
Effects of glyphosate spraying on morphological variables
Morphological variables of legumes seedlings were affected by the main factors C and G and by the interaction between them (Figure 3, page 42).
Vertical bars indicate 1 SD of the means. Different letters indicate significant differences provided by the Tukey test of the interaction factor (P< 0.05). Insert, results of ANOVA test: Main factors are Cover on which glyphosate was sprayed (C, at two levels C+and C-) and Spraying of glyphosate before sowing (G, at five levels G1, G30, G60, G90 or G0). Significant levels (SL) are indicated with an asterisk (*) when P<0.05 and with two asterisks (**) when P<0.01. Las barras verticales indican 1 desvío estándar de las medias. Letras distintas indican diferencias significativas según el test de Tukey (P< 0,05) para el factor interacción. Se insertaron los resultados del ANOVA: Los factores principales son Cobertura sobre la que se pulverizó glifosato (C, a dos niveles: C+y C-) y Pulverización de glifosato antes de la siembra (G, a cinco niveles: G1, G30, G60, G90 or G0). El nivel de significancia (SL) se indica con un asterisco (*) cuando P<0,05 y con dos asteriscos (**) cuando P<0,01.
Figura 3: Cantidad de tallos (a) y longitud de raíces (b) de Lotus tenuis, longitud de pecíolos (c) y longitud de raíces (d) de Trifolium repens, longitud de raíces of Paspalum dilatatum (e) y longitud de raíces de Festuca arundinacea (f) pulverizadas con glifosato sobre la vegetación pre-existente de L. multiflorum (C+; barras naranjas) o sobre suelo desnudo (C-; barras blancas) 1, 30, 60, 90 días antes de la siembra y el tratamiento control no pulverizado (G1, G30, G60, G90 y G0).
The number of stems of L. tenuis (Figure 3 a, page 42), the petiole length of T. repens (Figure 3 c, page 42) and the root lengths of both species (Figure 3 b and Figure 3 d, page 42) were significantly reduced under C+ at G1 and G30 respect to under C+ at G90 and G0 and under C- at any G level. The number of leaves per tiller and the leaf length of grasses were not affected by any treatments, but the root length of both grasses P. dilatatum and F. arundinacea was lower under C+ at G1 and G30 respect to under C+ at G60, G90 and G0 and under C- at any G level (Figure 3 e and Figure 3 f, page 42).
Discussion
Our results showed that glyphosate spraying on preexisting vegetation cover 1 to 60 days before sowing reduces seedlings emergence and growth of the forage legumes Lotus tenuis and Trifolium repens, and the forage grasses Paspalum dilatatum and Festuca arundinacea. The negative effect showed a threshold response because it prevailed during 30 or 60 days since glyphosate spraying until sowing, depending on the variable and the forage species.
We found that glyphosate greatly reduced seedling emergence and seedling growth when it was sprayed on pre-existing vegetation but not when it was sprayed on bare soil. Glyphosate mineralization rate in soil is much lower when it was previously absorbed by plants than when it is applied to bare soil 10,21. Herbicides trapped in plant residues provides protection against microbial degradation 21, so glyphosate degradation in soil following their absorption by plants is much slower, which increase it persistence from two to six times respect to directly sprayed on bare soil 10. Glyphosate or AMPA are released from tissues of treated plants to the rhizosphere zone and transfer to non-target plants via contact contamination 5,25,33. On the contrary, when glyphosate is sprayed on bare soil, it suffers a rapid microbial degradation or inactivation by sorption to the soil matrix 15.
Our results indicated that the reduction of seedling emergence when glyphosate was sprayed on pre-existing vegetation can be attributed to the toxic effect on germinated seeds, after rootlet appearance, because seed germination and seed viability were not affected. The mortality of germinated seeds occurred before seedling emergence, suggesting a greater susceptibility to the herbicide in younger tissues and seedlings of small size, susceptibility that decreases as the size or the state of development of the plants increases 20,35. Contrary to the studies that found reduction of seed germination when seeds germinated on boxes on filter paper and watered with glyphosate-water solutions 17,22, seed germination was not affected when glyphosate was sprayed on vegetation cover or bare soil, not directly over the seeds.
Seedlings that survived to glyphosate sprayed on the pre-existing vegetation, suffered a decline of its growth. Both legume and grass seedlings showed lower belowground biomass and root length, and seedlings of legumes also showed much lower aboveground biomass. These are indirect symptoms, mainly expressed as limitations to root functioning like absorption of water 38 and nutrients 25, which result in lower above and belowground growth. Moreover, our results suggest higher susceptibility to glyphosate of legumes than grasses, as species, and even biotypes of the same species, differ in their susceptibility to glyphosate 3,6.
Glyphosate persistence in soil is extremely variable as it is affected by soil properties, method of application, and environmental conditions, such as moisture and temperature 11. Our experiment showed that, in a typic Natracuol soil, the deleterious effect of glyphosate on seedling emergence and seedling growth had a threshold response pattern as it prevailed during at least 30 or 60 days between glyphosate spraying and sowing, with similar intensity along this period. Therefore, the recommendation of spraying glyphosate 7-30 days before seeding or overseeding pastures and forage species by no-tillage methods 1,2,37 would not be appropriate in order to avoid injury on seed emergence and seedling growth.
Conclusion
Our results suggest that the elapsed-time between glyphosate spraying on pre-existing vegetation of L. multiflorum and sowing of L. tenuis, T. repens, P. dilatatum and F. arundinacea seeds in argillic and natric soils must be longer than 60 days to avoid a reduction of seedling emergence and to maximize seedlings size and biomass. This finding is relevant for management purposes because of the prevalence of argilic and natric soils such as Argiudols and Natracuols in the Pampa region, where seeding or overseeding pastures and annual forage crops by no-tillage methods after glyphosate spraying is an extended practice.
