Implications for food safety of the uptake by tomato of 25 trace-elements from a phosphogypsum amended soil from SW Spain

Implications for food safety of the uptake by tomato of 25 trace-elements from a phosphogypsum amended soil from SW Spain Santiago Enamoradoa, José M. Abril a,∗, Antonio Delgado b, José L. Mása, Oliva Polvillo c, José M. Quintero b a Departamento de Física Aplicada I, Universidad de Sevilla, EPS, c/Virgen de África n◦ 7, D.P. 41011, Seville, Spain b Departamento de Ciencias Agroforestales, Universidad de Sevilla, ETSIA, Ctra. Utrera km. 1, D.P. 41013, Seville, Spain c Agricultural Research Service of the University of Seville (SIA-CITIUS), Ctra. Utrera Km. 1, D.P. 41013, Seville, Spain h i g h l i g h t s •Uptake of 25 trace-elements by tomato was studied in phosphogypsum-amended soils. •Cadmium had high transfer factors in tomato shoots (4.0) and fruits (1.5). •These two TF values of Cd increased to 5.7 and 2.5 after PG-inputs, respectively. •PG decreased the concentration of B, Cu, Sb, Cs, Ba, Tl and Th in fruits. •Assessment on trace elements in PG is needed to avoid exceeding legal limits in food. Keywords: Phosphogypsum Soil amendment Heavy metals Transfer factor Food safety a b s t r a c t Phosphogypsum (PG) has been usually applied as Ca-amendment to reclaim sodic soils such as those in the marshland area of Lebrija (SW Spain). This work aimed at the effects of PG amendments on the uptake of trace-elements by tomato and its implications for food safety. A completely randomized experiment was performed using a representative soil from Lebrija in a greenhouse involving six replicates and four PG treatments equivalent to 0, 20, 60, and 200 Mg ha−1. Soil-to-plant transfer factors (TFs) were determined for Be, B, Al, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, As, Se, Sr, Mo, Ag, Cd, Sb, Cs, Ba, Tl, Pb, Th and U. The highest TF in shoots was observed for Cd (4.0; 1.5 in fruits), its concentration being increased with increasing PG doses due to its content in this metal (2.1 mg Cd kg−1 PG). Phosphogypsum applying decreased the concentrations of Mn, Co and Cu in shoots; and of B, Cu, Sb, Cs, Ba, Tl and Th in fruits, however enhanced the accumulation of Se in fruits. Although Cd concentrations in tomato were below the maximum allowed levels in control pots (0 Mg PG ha−1), PG amendments above 60 Mg ha−1 exceeded such limits. 1. Introduction Phosphogypsum (PG) is a by-product of the industrial production of phosphoric acid, it being generated at a mass ratio of 5/1. Large quantities of PG are annually produced worldwide (about 170 million tons in 2006 [1]), most of them being stockpiled. For safety and economical reasons, there is a great interest in finding beneficial uses of PG, able to revalue and eliminate this waste. Use of PG in agriculture can be of particular interest because it is a Ca-source which can improve the functioning of sodic and acidic soils. Also, it can supply S and P to crops. Studies have been conducted, among Abbreviation: PG, phosphogypsum. ∗Corresponding author. Tel.: +34 954486473; fax: +34 954486436. E-mail address: [email protected] (J.M. Abril). others, on the effect of PG applying on the improvement of soil structure, salt balance and crop yield [2–4], control of soil erosion [5], availability of S and P [6], and N balance on reclaimed drained soils [7]. Thus, it can be supposed that agriculture could become the main worldwide sink of PG. Depending on the origin of the phosphate rock, concentrations of some heavy metals (e.g. Cd) and natural radionuclides (e.g. 226Ra) in PG can be higher than in agriculture soils by one order of magnitude [8,9]. Due to the associated potential hazards, some countries have developed regulations for the uses of PG, as the US-EPA 64 FR 5574 [10], or the Spanish Law of 19 July 2005 which allows its use as a soil amendment. The European Regulation No. 466/2001 establishes maximum levels of Cd, Pb and Hg in plant material and fruits for human consumption. Thus, some studies have been performed on the environmental impact and the radiological and alimentary safety in the agricultural use of PG [1,9,11–13]. On the other hand, it has been found that PG reduces the leaching of As, Cd, Tl, Pb, Zn and Ni in acid soils [14,15]. Thus, this by-product is considered for soil remediation practices in metal polluted soils. Phosphogypsum can affect plant nutrition either as a direct source of nutrients (Ca, S, P [6,16]) or varying soil geochemistry [15,17,18]. It can also enhance the concentrations in plant tissues of some hazardous elements, thus making the study of PG effects on soil-to-plant transfer of these elements a complex task of broad interest [1]. The agricultural use of PG in SW Spain is a relevant topic since the phosphoric acid industry located in Huelva generated a legacy of over 100 million tons of PG accumulated over four decades. The radionuclide and heavy metals content of this PG is well characterized [9,19]. Since the last sixties, this PG was crucial in the reclamation of the saline-sodic marshlands from SW Spain [4,20–22] where it was applied as Ca-amendment at a usual rate of 25 Mg ha−1(wet weight, after being sun-dried, and with typical remaining water content of 20%). This amendment was essential to achieve a reduction of Na saturation and improvements of physical properties in soil [6,7,23]. On the other hand, there are evidences of significant enrichment of 226Ra in the surface horizon of reclaimed soils plots relative to the deeper horizons and Cd in tomatoes produced in the area with higher concentrations than those found in tomatoes from other origins in Spain [1,9]. These latest are long term effects, evident after three decades of PG amendments. In farm soils under traditional practices, there may be different sources of Cd and other harmful pollutants (e.g. phosphated fertilizers, organic and inorganic soil amendments), and their cycling and interactions in agro-ecosystems are still poorly understood. Thus, it seems advisable complementing the previous works with an experimental study conducted under controlled conditions. The aim of the present paper was to carry out such experimental study on trace-elements uptake by plant as being affected by different PG doses with a view to state its implications for food safety. Special focus was put on the soil-to-plant transfer of 25 trace-elements (Be, B, Al, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, As, Se, Sr, Mo, Ag, Cd, Sb, Cs, Ba, Tl, Pb, Th and U) present in PG. To this end, a well characterized PG source, a representative soil from the area of Marismas of Lebrija, and a crop of great economical and social importance in SW Spain (industrial tomato) had been used. Results have been analyzed in terms of legal regulations on their potential health effects due to the consumption of tomato fruits by humans and of plant debris by livestock. Results obtained, beside their regional interest, can provide information and tools to study the potential effects of the agricultural use of PG and other hazard products on the environment and food safety. 2. Experimental 2.1. Area of study, soil and crop The area known as Marismas de Lebrija accounts for most of 40,000 ha of originally saline-sodic soils in SW Spain. Reclamation of this area started at the end of the seventies of last century, with the installation of a tile-drain system and the application of PG amendments. After reclamation, most of the soils in the area can be classified as Aeric Endoaquepts [24]. More information about the area is available elsewhere [6,21,23]. The main crops are cotton, sugar beet, cereals (wheat and corn), and some vegetables. In recent years, industrial tomato is progressively replacing cotton and sugar beet in many farms. The area is also important by its extensive livestock (ovine, bovine, poultry and horse) which are fed with grass and plant debris produced locally. Surface soil (ca. 0–30 cm) was collected in January 2005 from ten points randomly selected throughout the area (6 ha) of a commercial farm located at 37◦1.2N, 6◦7.4W. This farm had received during the last decades up to six PG amendments at typical doses of 20–25 Mg ha−1[9]. General soil properties in the 0–30 cm horizon were characterized by Hurtado et al. [4]. Texture analysis provided 85, 372 and 543 g kg−1of sand, silt and clay, respectively. Organic carbon and calcium carbonate equivalent (CCE) were 6.4 and 235 g kg−1, respectively; and its cation exchange capacity (CEC) was 32 cmolckg−1. Finally, the pH determined in the 1:5 soil extract was 8.5. Soil was air-dried, homogenized and ground to pass a 2- mm screen, and 11.6 kg were added to each 15-L pot along with the perlite (0.8 kg). PG was collected from an inactive stack in Huelva (37◦15.3N, 6◦54W) and processed as the soil. 2.2. Experimental design and samples Industrial tomato (Solanum lycopersicum L.) was grown under greenhouse conditions in pots using as substrate the soil described above, mixed with inert material (1/3, v/v ratio of expanded perlite) to ensure a viable soil aeration and hydraulic conductivity. A completely randomized experiment was performed involving six replications and four treatments: C (farm soil without any new PG addition), T1, T2 and T3 (as C, but with 1, 3 and 10 usual doses of PG, respectively). The addition of 60 g of PG to the previous substrate in each 15 L pot will represent one dose, which is the equivalent to the usual rate applied to soil in the field, 25 Mg ha−1, considering that this amendment is mixed with soil to depth of 40 cm. As fertilizer, each pot received 60 g of Osmocote®supplied by Scotts Espa˜ na S.A. (Tarragona, Spain) with a nutrient content of 18% N, 10% P2O5, and 11% K2O. Beside this, each plant received 1 L of a liquid foliar fertilizer Ca(NO3)2during fruit development stage to avoid the nutritional disorder blossom-end-rot. Plants were transferred to pots after 1–1.5 months at seedbed and allowed to produce fruits. Volumes of irrigation water during the experiment were monitored, accounting for 54.6 L for each pot. Drain water samples were collected for each irrigation episode in clean trays (previously washed with 2% HNO3analysis quality diluted in 18 Mdeionised water). Overall, drained volumes accounted for 5% of the irrigation water. The complete aerial parts were collected after 111 days, in June 2005, and separated into fruits and shoots (stems and leaves) parts. Aerial parts were dried at 36 ◦C for 48h, and after that cut into small pieces and ground. Fruits were ground using a juice mixer, dried for 5 days as described for shoots, homogenized, and ground again. Measurements of fresh and dried weights allowed for the determination of the water content in the organic samples. At the end of the experiment, soil samples were recovered from each pot and carefully separated from roots and the inert material. After that, the soils were dried in clean steel trays and representative samples were submitted for analysis. The concentrations of 25 trace-elements, namely Be, B, Al, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, As, Se, Sr, Mo, Ag, Cd, Sb, Cs, Ba, Tl, Pb, Th and U were determined in substrate, plant, and drainage water after the experiment. Also, these elements were determined in the PG, initial soil, perlite, fertilizers and a representative sample of irrigation water to quantify their inputs into the system. 2.3. Sample preparation and ICP-MS analysis Irrigation and drain water samples were filtered through 0.22 ␮m Millipore©filters to measure analytes concentrations in the dissolved fraction. A new dilution was required in most of the samples due to the high value of the total dissolved solids (routinely higher than 0.1%). In all the cases, the sample matrix was adapted to 1% HNO3. Several digestion procedures were tested for biological matricesusingcertified referencematerial,including tomato leaves(SRM 1573a), as reported elsewhere [25]. The one selected for this work used 0.5 g of sample with 10 mL of nitric acid (69% HNO3) of Suprapure®grade (Merck, Darmstadt, Germany). For digestions in closed-vessels, a microwave system Multiwave 3000 (Anton Paar, Graz, Austria) with rotor HF100 and software version v1.52 was used. The same digestion method was selected for the rest of solid samples (soils, PG and Osmocote fertilizer). It is worth noting that the selected method applied to soil samples provides a reasonable estimation of leachable/total contents of elements [26], being more representative of the bio-available fraction, and eliminating bias introduced by the variable amount of non-reactive residual material [27]. Finally, the foliar fertilizer was prepared with the adapted US-EPA 3015A method after evaporating 500 mL to 100 mL in a Teflon beaker on a hot plate at T< 95 ◦C; 1 mL of concentrated suprapure HNO3was added to 9 mL of the remaining solution for digestion. After digestion, samples were cooled down to room temperature. The aliquots were then diluted with 18 Mdeionised water until reaching a final nitric acid concentration of ca. 1%, and filtered through 0.45 ␮m pore size Millipore filter. An Inductively Coupled Plasma Mass Spectrometry system Thermo Elemental ICP-MS X7 (Thermo Fisher, Cambridge, UK) with quadrupole mass analyzer, multichannel detector (Pulse Counting and Analog Methods), auto sampler ASX-500 (CETAC, Omaha, NE, USA) and software Plasma Lab version v4.5, was used for this work. The instrument, located at the Servicio de Investigación Agraria laboratory (University of Seville, Spain), was used with a concentric Meinhard type glass nebulizer, a silica impact bead spray chamber, cooled to 3 ◦C by a Peltier cooler, and a standard silica torch. Standard nickel sample and skimmer cones were used. The internal standard solution was added online by a “Y” connection in the pipe where the sample is aspired by the peristaltic pump. Details on the instrumental settings, correction equations, quality controls and sample lists can be found in Enamorado et al. [25]. Methods were tested through the analysis of different certified reference materials: CRM-TMDW (water), SRM 1573a (tomato leaves) and SRM 2711 (Montana Soil). 2.4. Calculation of soil-to-plant transfer ratios The soil-to-plant transfer factor (TF) of any given element will be operationally defined as the ratio between the element concentration in plant (shoots or fruits) and the corresponding concentration in substrate (without perlite particles) at the end of the experiment, given both concentrations on a dry weight basis. This operational definition includes the use of traceelement concentrations determined after the microwave-assisted acid digestion methods described above for sample mineralization. 2.5. Statistical analysis Analysis of variance to see the effect of the different treatments on the concentrations of studied trace-elements in plant tissues, soil, and drainage waters was performed by using the general linear model in Statgraphics plus 5.1 software. Homogeneity of variance was assessed by means of the Cochran, Levene, and Bartlett tests; the last of which is sensitive to departures from normality. The Tukey test at 95% CL was used for comparison between means associated to four treatments (C, T1, T2, T3) with six replications for shoots and tomato fruits. For soil samples we used only four replications, and three for drainage waters. Table 1 Method Detection Limits (MDL) determined from laboratory blank solutions. Element m/zMDL filtered water (␮g L−1) MDL organic/solids (mg kg−1) Be 9 0.008 0.010 B 11 0.17 0.76 Al 27 0.16 2.9 Ti 47 0.1 2.5 V[1] 51 0.12 V 51 0.008 0.12 Cr 52 0.03 0.055 Mn 55 0.005 0.033 Fe 56 1.3 10.2 Co 59 0.002 0.012 Ni 60 0.11 0.15 Cu 63 0.01 0.33 Cu 65 0.02 0.50 Zn 66 0.06 1.4 As 75 0.01 0.020 As [2] 75 0.27 Se [3] 82 0.87 Se 82 0.1 0.88 Sr 88 0.004 0.045 Mo 95 0.006 0.26 Ag 107 0.001 0.081 Cd [4] 111 0.005 0.083 Sb 123 0.003 0.005 Cs 133 0.003 0.014 Ba 137 0.017 0.059 Tl 205 0.001 0.003 Pb [5] 208 0.005 0.024 Th 232 0.01 0.005 U 238 0.001 0.002 m/z, mass to charge ratio. Correction equations: [1] 51V: -0.35252Cr-3.12753Cr; [2] 75As: -0.03182Kr- 3.132277ArCl; [3] 82Se: -1.00183Kr; [4] 111Cd: -0.764106Cd-1.073108Cd; [5] 208Pb: 1.00206Pb + 1.00207Pb 3. Results 3.1. Reagent blanks, Method Detection Limits and recovery of certified samples Detection limits for each sample preparation method (MDL) were determined from reagent blanks by using the US-EPA 200.8 definitions (Table 1). Each reagent blank was prepared using the same amount and mixture of acids and following the same experimental procedure than the unknown samples. When reagent blanks were treated as unknown samples with subtraction of the calibration blank, only Al (2.4–5.5 ␮g L−1), an airborne pollutant, was significantly above the instrument detection limits. Table 2 shows the measurements of three reference certified materials (three replicates each) with the sample preparation procedures used in this work. The average recovery for filtered waters was (99 ±5)% (ranging from 89.7% for Ni to 113% for Fe; Fig. 1). In tomato leaves, when excluding anomalous results in Ag, Cs and Sr (non certified concentrations), the average recovery was (92 ±17)% (ranging from 70% for Cu to 133 ±20% for As; Fig. 1). In soils, recoveries under 50% were obtained for Ti (30%), Ba (39%) and Sb (47 ±5%) (Fig. 1). Anomalous high recovery was found for Ni (180%), likely due to the isobaric interference of 44Ca +16O, which was not properly addressed through correction equations or other methods. This isobaric interference is of particular concern in Ca rich matrices such as those derived from the soil used. Then, the reported values for Ni in soils and their corresponding TF have to be considered on a semiquantitative basis. For the rest of elements, the averaged recovery was (89 ±19)% (ranging from 54% for Al to 116% for Pb). As expected, the acid digestion of soils with nitric acid provided a pseudo-total recovery. Table 2 Measured concentrations in certified materials following the sample preparation procedures (filtered waters and acid digestion for solids - 0.5 g of sample with 10 mL of HNO3). Element m/zCRM-TMDWtraceelements in water SRM 1573a tomato leaves SRM 2711 montana soil Measured Certified#Measured Certified Measured Certified ␮g L−1mg kg−1 Be 9 19.2 ±0.2 20 0.0127 ±0.0022 B 11 30.2 ±0.4 Al 27 120.6 ±1.3 120 567 ±23 598 ±12 3.53 ±0.13+6.53 ±0.09+ Ti 49 24.0 ±1.7 0.092 ±0.006+0.306 ±0.023+ V 51 0.79 ±0.04 0.835 ±0.010 61.2 ±0.8 81.6 ±2.9 Cr 52 19.3 ±0.1 20 1.65 ±0.10 1.99 ±0.06 35 ±3 47* Mn 55 38.5 ±0.3 40 219 ±6 246 ±8 708 ±26 638 ±28 Fe 56 113 ±2 100 264 ±10 368 ±7 2.66 ±0.09 2.89 ±0.06+ Co 59 24.6 ±0.2 25 0.696 ±0.019 0.57 ±0.02 9.80 ±0.03 10* Ni 60 53.8 ±0.6 60 1.64 ±0.06 1.590 ±0.07 37.1 ±0.6 20.6 ±1.1 Cu 65 18.5 ±0.2 20 3.29 ±0.14 4.70 ±0.14 106 ±3 114 ±2 Zn 66 25.9 ±0.9 30.90 ±0.7 322 ±7 350 ±5 As 75 80.6 ±0.2 80 0.149 ±0.022 0.112 ±0.004 59 ±3 105 ±8 Se 82 9.5 ±0.2 10 0.054 ±0.003 1.3 ±1.2 1.52 ±0.14 Sr 88 264 ±2 250 76.9 ±1.8 85*260.0 ±2.4 245.3 ±0.7 Mo 98 96.8 ±1.1 100 0.362 ±0.015 0.96 ±0.14 1.6* Ag 107 2.00 ±0.03 2.0 0.029 ±0.018 0.017*5.0 ±0.5 4.63 ±0.39 Cd 111 10.30 ±0.16 10 1.47 ±0.07 1.52 ±0.04 46.3 ±2.1 41.70 ±0.25 Sb 123 10.40 ±0.07 10 0.063 ±0.006 9.1 ±0.4 19.4 ±1.8 Cs 133 0.017 ±0.0020 0.053*5.25 ±0.06 6.1* Ba 137 50.00 ±0.04 50 55.3 ±2.0 63*285 ±7 730 ±40 Tl 205 10.00 ±0.01 10 0.041 ±0.026 2.37 ±0.07 2.47 ±0.15 Pb 208 36.2 ±0.3 40 0.548 ±0.0255 1350 ±30 1162 ±31 Th 232 0.095 ±0.004 0.120*14.6 ±0.4 14* U 238 10.10 ±0.18 10 0.0293 ±0.0021 0.035*2.05 ±0.08 2.6* Measured values are reported as the mean and standard deviation (n= 3). Non reported values were under their respective MDL (Table 1). #Certified values provided with ±0.5% uncertainties. *Reference values. +Values given as percentage. 3.2. Results on multi-elemental analysis The measured element concentrations in soil, irrigation water, PG, perlite, and fertilizers used in the experiment are summarized in Table 3. Concentrations of B, Cd and U in PG and of Mo in the perlite were roughly one order of magnitude higher than those found in soil. Table 4 shows results on water content determinations in shoots and tomato fruits. Table 5 summarizes the measured element concentrations in soil, shoots (aerial non generative part of plants) and tomato fruits (in mg kg−1dry mass) at the end of the experiment, and the total amounts lixiviated through the drainage waters (␮g), for the four treatments. There was not always an increased concentration of trace-elements with increasing PG doses. In shoots and fruits, Cd was the only element which significantly increased at increased PG doses. In other cases, such as Cu, increased PG doses resulted in decreased concentrations in plant tissues (Table 5). 3.3. Soil-to-plant transfer factors Transfer factors (TF) for shoots and tomato fruits can be evaluated, for all the elements and treatments, from raw data in Table 5. For the sake of brevity only TF for control pots are reported in Fig. 2 and Table 6. They varied up to three orders of magnitude for 0 50 100 150 200 UThPbTlBaCsSbCdAgMoSrSeAsZnCuNiCoFeMnCrVTiAlBBe Recovery (%) SRM 1573a Tomato Leaves SRM 2711 Montana Soil CRM-TMDW Cerfied water Fig. 1. Summary of recoveries for certified reference materials of different nature obtained with the sample preparation methods used in this work: tomato leaves, soil and water. Results represent the average (n= 3) and standard deviation. Table 3 Element concentration in soil, irrigation water, PG, inert material and fertilizers used in the experiment. Element Soil*Phosphogypsum Osmocote Inert material Irrigation water Ca(NO3)2 mg kg−1␮g L−1 Be 1.92 ±0.04 0.42 ±0.05 0.11 ±0.01 N.D. 0.005 ±0.0008 0.094 ±0.005 B 49 ±6 3120 ±160 88 ±64 N.D. 6.7 ±0.2 78 ±1 Al 36,100 ±1300 3040 ±30 350 ±20 64 ±156±2 N.D. Ti 113 ±21 39 ±6 6.0 ±0.7 10.0 ±1.3 0.018 ±0.002 1.8 ±0.6 V 81 ±2 N.D. 5.4 ±1.2 0.88 ±0.3 0.89 ±0.03 1.1 ±0.1 Cr 66 ±6 18.3 ±0.3 8.5 ±0.2 N.D. 1.01 ±0.02 0.41 ±0.02 Mn 782 ±29 0.28 ±0.03 290 ±12 3.6 ±0.4 0.31 ±0.01 52 ±1 Fe 32,600 ±1100 366 ±9 1700 ±50 160 ±30 35 ±3 N.D. Co 14.1 ±0.2 1.47 ±0.01 6.8 ±0.6 0.023 ±0.001 0.100 ±0.004 13.2 ±0.2 Ni 41 ±112±2 15.0 ±1.4 N.D. N.D. 112±2 Cu 27 ±1 5.2 ±0.5 6.9 ±0.6 0.15 ±0.03 7.9 ±0.5 13.0 ±0.2 Zn 89 ±2 11 ±2 120 ±8 3.0 ±0.6 214 ±2 N.D. As 12.8 ±0.6 2.6 ±0.2 0.62 ±0.02 0.15 ±0.05 0.48 ±0.03 0.25 ±0.05 Se 0.9 ±0.8 2.8 ±0.7 0.8 ±0.1 0.8 ±0.1 0.37 ±0.10 0.64 ±0.06 Sr 289 ±11 538 ±2 520 ±30 0.47 ±0.12 64 ±1 720 ±2 Mo 0.16 ±0.05 0.88 ±0.15 104 ±10 2.5 ±0.6 0.43 ±0.01 0.51 ±0.05 Ag 0.16 ±0.02 0.20 ±0.01 0.022 ±0.002 100 ±8 0.0011 ±0.0001 0.017 ±0.005 Cd 0.19 ±0.01 2.14 ±0.07 0.31 ±0.03 0.022 ±0.002 N.D. 0.650 ±0.005 Sb 0.027 ±0.010 0.411 ±0.013 0.117 ±0.003 0.23 ±0.01 0.193 ±0.012 0.106 ±0.005 Cs 6.11 ±0.17 0.020 ±0.006 0.16 ±0.01 N.D. 0.0070 ±0.0001 0.072 ±0.005 Ba 179 ±8 91 ±2 27 ±1 N.D. 13.2 ±0.8 27 ±1 Tl 0.38 ±0.02 0.025 ±0.005 0.070 ±0.003 27 ±1 0.0090 ±0.0005 N.D. Pb 36 ±4 0.030 ±0.002 0.42 ±0.01 0.0037 ±0.00 1.62 ±0.02 2.50 ±0.13 Th 7.86 ±0.22 2.42 ±0.02 2.7 ±0.1 0.42 ±0.01 0.003 ±0.001 0.09 ±0.01 U 1.72 ±0.07 16.0 ±0.2 3.7 ±0.2 N.D. 0.075 ±0.002 0.070 ±0.002 Amounts added at each pot: 11.6 kg of soil; 0.060 kg of Osmocote; 0.060 kg for each PG dose (0, 1, 3 and 10 doses for C, T1, T2and T3, respectively); 0.80 kg of inert material, and 54.6 L of irrigation water. 1 L of Ca(NO3)2as foliar fertilizer at pre-fruit stage. N.D., non detected (under MDL) *Single measurement of a representative soil sample. Associated uncertainties incorporates variability in the digestion yield of certified soil samples (Table 2). Table 4 Water content (% of fresh weight) in shoots (aerial non generative part of plants) and tomato fruits for the different treatments. C T1 T2 T3 Shoots 78.3 ±0.2 77.4 ±0.3 77.1 ±0.4 77.0 ±0.2 Fruits 86.2 ±0.2 85.8 ±0.3 85.5 ±0.3 85.4 ±0.2 Mean and standard deviation (n= 6). different elements, the highest values being found for Mo and Cd in both fruits and shoots (Fig. 2). For most of the studied elements present over their MDL, transfer to shoots was higher than to fruits, except for Mo. The so obtained values in our experimental setting could be partially perturbed by the foliar application. Taking into account the concentrations of tested elements in Ca(NO3)2 fertilizer, the used amount (Table 3), and the dry masses of shoots and fruits (49.1 ±0.9 and 89.5 ±1.5 g, respectively – mean and standard deviation for control pots), the maximum contribution of the foliar fertilizer to the TF for shoots was below 1.5% for most of the studied elements. Exceptions were Co and Ni, whose TF were increased up to 12% and 4% by this means, respectively; and, to a lesser extent, B and Sr, with contributions to TFs around 3%. The foliar fertilizer was applied at the time of the appearance of the first fruits, and owing to their concentrations, the TF for Co and Ni in fruits should be also handled with care. Transfer factors have been published for a large variety of elements, soils and crops (Table 6 shows some of them for the sake of comparison). For most of the elements, observed TFs were comparable to those reported in the bibliography, at least in order of magnitude. It is expected that TF will depend on the state and geochemistry of the element in soil, on the plant type, and on the involved physiological processes. Thus, present results have the -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 Th Al V Cs Fe Be U Pb Cr Co Mn As Ba Tl Ni Ti Zn Cu Sr B Sb Ag Se Mo Cd Log[TF] Shoots Fruits Fig. 2. Log of soil-to-plant transfer factors (as operatively defined in this work) for both shoots and fruits. Table 5 Element concentrations in soil, shoots (aerial non generative part of plants) and tomato fruits (in mg kg−1d.w.), and total amounts lixiviated through the drainage waters (␮g). C T1 T2 T3 Be Soil 1.77 ±0.11 (a) 1.6 ±0.2 (a) 1.76 ±0.06 (a) 1.75 ±0.07 (a) Shoots 0.10 ±0.03 (a) 0.06 ±0.03 (a) 0.06 ±0.02 (a) 0.060 ±0.013 (a) Fruits N.M. (a) N.M. (a) N.M. (a) N.M. (a) Drainage Water 0.52 ±0.05 (a) 0.20 ±0.02 (c) 0.43 ±0.06 (b) 0.22 ±0.02 (a) B Soil 55 ±8 (a) 54 ±11 (a) 53 ±2 (a) 51 ±10 (a) Shoots 29 ±4 (ab) 25 ±4 (b) 30 ±3 (a) 30 ±3 (ab) Fruits 18.3 ±1.1 (a) 17 ±2 (ab) 15.8 ±1.1 (ab) 15 ±2 (b) Drainage Water 1114 ±17 (a) 877 ±12 (c) 1058 ±14 (b) 567 ±10 (d) Al Soil 40,000 ±4000 (a) 39,000 ±6000 (a) 38,900 ±1100 (a) 38,000 ±6000 (a) Shoots 1800 ±600 (a) 1400 ±800 (a) 1400 ±700 (a) 1200 ±200 (a) Fruits 700 ±200 (a) 600 ±300 (a) 450 ±230 (a) 520 ±110 (a) Drainage Water 183 ±4 (a) 27 ±3 (b) 190 ±16 (a) 40.2 ±1.2 (b) Ti Soil 110 ±70 (a) 130 ±80 (a) 113.3 ±1.8 (a) 100 ±80 (a) Shoots 42 ±11 (a) 35 ±13 (a) 33 ±10 (a) 32 ±8 (a) Fruits 2.9 ±0.4 (a) 2.7 ±0.4 (a) 2.5 ±0.3 (a) 2.5 ±0.4 (a) Drainage Water 15.1 ±1.9 (a) 3.8 ±0.9 (c) 10.1 ±1.2 (b) 5.2 ±0.9 (c) V Soil 69 ±12 (a) 67 ±16 (a) 63 ±4 (a) 64 ±19 (a) Shoots 3.2 ±1.3 (a) 3.1 ±1.2 (a) 3.1 ±1.4 (a) 2.3 ±0.8 (a) FruitsaN.M. (a) N.M. (a) N.M. (a) N.M. (a) Drainage Water 72.8 ±1.5 (a) 64.8 ±1.1 (b) 76 ±3 (a) 45.6 ±1.1 (c) Cr Soil 59 ±4 (a) 59 ±8 (a) 58.6 ±1.7 (a) 58 ±6 (a) Shoots 4.4 ±1.4 (a) 3.1 ±1.0 (a) 3.0 ±0.6 (a) 3.2 ±0.3 (a) Fruits 0.70 ±0.19 (a) 0.52 ±0.09 (a) 0.60 ±0.11 (a) 0.70 ±0.07 (a) Drainage Water 7.1 ±0.4 (b) 6.1 ±0.2 (c) 8.8 ±0.3 (a) 4.89 ±0.14 (d) Mn Soil 850 ±30 (a) 850 ±40 (a) 840 ±20 (a) 834 ±17 (a) Shoots 74 ±13 (a) 51 ±10 (b) 55 ±9 (b) 58 ±4 (ab) Fruits 19.4 ±1.2 (a) 19 ±2 (a) 17.5 ±1.6 (a) 18 ±2 (a) Drainage Water 12.65 ±0.13 (b) 4.81 ±0.11 (c) 14.5 ±0.5 (a) 3.99 ±0.09 (d) Fe Soil 32,600 ±500 (a) 34,400 ±1300 (a) 32,000 ±900 (a) 32,200 ±1100 (a) Shoots 1600 ±300 (a) 1200 ±700 (a) 1100 ±300 (a) 1120 ±170 (a) Fruits 63 ±9 (ab) 50 ±30 (b) 80 ±30 (ab) 88 ±18 (a) Drainage Water 780 ±50 (a) 320 ±30 (b) 380 ±40 (b) 77 ±14 (c) Co Soil 13.6 ±0.3 (b) 14.4 ±0.4 (a) 13.64 ±0.15 (b) 13.61 ±0.09 (b) Shoots 1.06 ±0.12 (a) 0.69 ±0.19 (b) 0.75 ±0.12 (b) 0.79 ±0.07 (b) Fruits 0.023 ±0.004 (a) 0.025 ±0.011 (a) 0.032 ±0.013 (a) 0.037 ±0.013 (a) Drainage Water 18.0 ±0.7 (c) 26.6 ±0.3 (a) 23.1 ±0.5 (b) 9.13 ±0.15 (d) Ni Soil 36.7 ±1.2 (ab) 39.3 ±1.7 (a) 36.7 ±0.5 (ab) 36.0 ±1.2 (b) Shoots 4.6 ±0.7 (ab) 3.4 ±0.6 (b) 5.3 ±1.7 (a) 3.9 ±0.2 (ab) FruitsaN.M. (a) N.M. (a) N.M. (a) N.M. (a) Drainage Water 59 ±7 (a) 50 ±5 (ab) 39 ±8 (b) 22 ±4 (c) Cu Soil 26.9 ±0.5 (b) 28.8 ±0.8 (a) 27.3 ±1.3 (ab) 27.6 ±0.8 (ab) Shoots 11.9 ±0.8 (a) 10.4 ±1.2 (a) 8.4 ±1.1 (b) 6.8 ±0.5 (b) Fruits 8.4 ±1.3 (a) 9.5 ±1.9 (a) 7.5 ±1.3 (ab) 5.8 ±1.2 (b) Drainage Water 83.3 ±1.5 (b) 60.8 ±0.8 (c) 102.8 ±1.5 (a) 35.7 ±0.5 (d) Zn Soil 72 ±3 (a) 75.3 ±1.1 (a) 75 ±2 (a) 78 ±10 (a) Shoots 32 ±4 (a) 24 ±3 (a) 25 ±4 (a) 27 ±9 (a) Fruits 31 ±3 (a) 32 ±2 (a) 28 ±3 (a) 27 ±5 (a) Drainage Water 334 ±10 (c) 5084 ±80 (a) 994 ±18 (b) 379 ±9 (c) As Soil 10.7 ±0.4 (a) 11.1 ±0.5 (a) 10.71 ±0.18 (a) 10.8 ±0.4 (a) Shoots 1.0 ±0.2 (a) 0.8 ±0.3 (a) 1.0 ±0.21 (a) 0.97 ±0.15 (a) FruitsaN.M. (a) N.M. (a) N.M. (a) N.M. (a) Drainage Water 26.7±0.5 (b) 27.2 ±0.9 (b) 33.3 ±1.2 (a) 18.8 ±0.6 (c) Se Soil 1.0 ±0.2 (a) 1.1 ±0.4 (a) 1.0 ±0.2 (a) 1.1 ±0.2 (a) Shoots 1.1 ±0.3 (a) 0.96 ±0.11 (a) 1.3 ±0.2 (a) 1.3 ±0.3 (a) Fruits 0.27 ±0.16 (b) 0.32 ±0.10 (b) 0.53 ±0.07 (ab) 0.7 ±0.2 (a) Drainage Water 30.5 ±1.7 (c) 37.8 ±2 (b) 49 ±2 (a) 19.1 ±1.4 (d) Sr Soil 242 ±8 (a) 257 ±17 (a) 246 ±6 (a) 250 ±10 (a) Shoots 266 ±7 (b) 237 ±7 (c) 273 ±10 (ab) 286 ±8 (a) Fruits 5.9 ±1.1 (a) 4.9 ±0.5 (a) 5.3 ±0.4 (a) 5.8 ±0.4 (a) Drainage Water 17,300 ±200 (b) 16,600 ±130 (c) 20,600 ±200 (a) 10,130 ±70 (d) Mo Soil 0.65 ±0.11 (a) 0.71 ±0.14 (a) 0.74 ±0.06 (a) 0.87 ±0.15 (a) Shoots 1.3 ±0.2 (a) 1.3 ±0.3 (a) 1.2 ±0.2 (a) 1.14 ±0.16 (a) Fruits 2.5 ±0.3 (a) 2.5 ±0.2 (a) 2.47 ±0.13 (a) 2.2 ±0.2 (a) Drainage Water 38.4 ±0.8 (c) 128 ±3 (a) 87.7 ±0.8 (b) 19.2 ±0.3 (d) Table 5 (Continued) C T1 T2 T3 Ag Soil 0.122 ±0.013 (a) 0.145 ±0.018 (a) 0.13 ±0.03 (a) 0.131 ±0.008 (a) Shoots 0.14 ±0.02 (b) 0.118 ±0.015 (b) 0.14 ±0.03 (b) 0.34 ±0.07 (a) Fruitsa0.20 ±0.05 (a) 0.23 ±0.06 (a) 0.19 ±0.04 (a) 0.20 ±0.07 (a) Drainage Water 0.11 ±0.04 (ab) 0.14 ±0.03 (a) 0.08 ±0.04 (bc) 0.028 ±0.011 (c) Cd Soil 0.19 ±0.02 (a) 0.21 ±0.02 (a) 0.209 ±0.018 (a) 0.23 ±0.02 (a) Shoots 0.75 ±0.09 (c) 0.81 ±0.06 (c) 1.08 ±0.18 (b) 1.32 ±0.12 (a) Fruits 0.278 ±0.019 (c) 0.35 ±0.03 (bc) 0.44 ±0.05 (b) 0.57 ±0.10 (a) Drainage Water 8.3 ±0.3 (c) 17.1 ±0.3 (b) 24.4 ±0.2 (a) 3.00 ±0.06 (d) Sb Soil 0.07 ±0.11 (a) 0.07 ±0.07 (a) 0.04 ±0.06 (a) 0.10 ±0.08 (a) Shoots 0.075 ±0.013 (a) 0.052 ±0.014 (a) 0.072 ±0.011 (a) 0.080 ±0.011 (a) Fruits 0.06 ±0.05 (a) N.M. (b) N.M. (b) N.M. (b) Drainage Water 6.62 ±0.07 (c) 8.74 ±0.18 (a) 7.67 ±0.19 (b) 3.02 ±0.09 (d) Cs Soil 5.7 ±0.6 (a) 5.9 ±0.6 (a) 5.60 ±0.05 (a) 5.4 ±0.8 (a) Shoots 0.28 ±0.10 (a) 0.22 ±0.14 (a) 0.23 ±0.12 (a) 0.19 ±0.04 (a) Fruits 0.8 ±0.7 (a) 0.002 ±0.004 (b) N.M. (b) N.M. (b) Drainage Water 1.37 ±0.06 (b) 0.83 ±0.08 (c) 1.74 ±0.18 (a) 0.56 ±0.06 (d) Ba Soil 159 ±19 (a) 162 ±19 (a) 154 ±10 (a) 150 ±30 (a) Shoots 19 ±4 (a) 16 ±4 (a) 17 ±3 (a) 16.8 ±1.2 (a) Fruits 1.1 ±0.4 (a) 1.0 ±0.4 (ab) 0.5 ±0.3 (bc) 0.32 ±0.17 (c) Drainage Water 386 ±6 (c) 401 ±4 (b) 435 ±5 (a) 217.3 ±1.5 (d) Tl Soil 0.32 ±0.04 (a) 0.33 ±0.04 (a) 0.313 ±0.014 (a) 0.31 ±0.05 (a) Shoots 0.039 ±0.014 (a) 0.037 ±0.014 (a) 0.033 ±0.009 (a) 0.028 ±0.005 (a) Fruitsa0.005 ±0.003 (a) N.M. (b) N.M. (b) N.M. (b) Drainage Water 0.284 ±0.019 (c) 1.38 ±0.03 (a) 0.81 ±0.03 (b) 0.138 ±0.015 (d) Pb Soil 32.3 ±0.7 (a) 30.47 ±0.13 (b) 30.81 ±0.10 (ab) 30.5 ±1.2 (ab) Shoots 2.4 ±0.7 (a) 1.7 ±0.7 (a) 2.0 ±0.7 (a) 2.3 ±0.2 (a) Fruits 0.14 ±0.11 (a) 0.14 ±0.03 (a) 0.04 ±0.03 (b) 0.054 ±0.019 (ab) Drainage Water 20.3 ±0.6 (c) 85.8 ±0.8 (a) 51.7 ±1.0 (b) 5.11 ±0.09 (d) Th Soil 9.92 ±0.19 (a) 10.1 ±0.12 (a) 9.84 ±0.11 (a) 9.71 ±0.19 (a) Shoots 0.27 ±0.11 (a) 0.24 ±0.13 (a) 0.19 ±0.11 (a) 0.17 ±0.03 (a) Fruits 0.18 ±0.09 (a) 0.024 ±0.019 (b) 0.03 ±0.05 (b) N.M. (b) Drainage Water 3.45 ±0.18 (c) 3.99 ±0.07 (b) 4.68 ±0.013 (a) 2.48 ±0.06 (d) U Soil 1.63 ±0.12 (a) 1.64 ±0.06 (a) 1.7 ±0.2 (a) 1.82 ±0.11 (a) Shoots 0.097 ±0.014 (a) 0.06 ±0.02 (c) 0.064 ±0.010 (bc) 0.089 ±0.014 (ab) FruitsaN.M. (a) N.M. (a) N.M. (a) N.M. (a) Drainage Water 33.1 ±0.5 (c) 36.4 ±0.4 (b) 45.5 ±0.04 (a) 28.7 ±0.2 (d) Concentrations correspond to mean and standard deviation of mean (n = 4 for soils, n = 6 for biological matrices and n = 3 for drain water samples).). Means followed by the same letter within a column and an element are not significantly different (Tukey’s test at the 95% CL). N.M. < MDL. Treatments C, T1, T2and T3as defined in the text. value of being representative of the studied crop and soil, both of great local social and economical relevance. 4. Discussion 4.1. TF and changes in element inventories in control pots Differences found in TF between elements of the same group, such as Zn and Cd, Fe and Ni or Sr and Ba can be explained by their geochemistry, particularly by their activity in the soil solution, and by the efficiency of uptake mechanisms by plants. One of the most extreme cases is Fe, an abundant element in soil, which however is present at very low concentrations in soil solution (usually lower than 6 ×10−19 mg L−1at the pH of the studied soil [31]) due to the low solubility of minerals containing it in basic soils. This low solubility restricts its uptake by plants. Depending on the mobilization strategies of plants roots, such as acidification, or the exudation of organic compounds with complexing capacity, which can be element-specific, the TF of metals could vary obviously. On the contrary, elements with increased solubility at basic soil pH, such as Mo [32], show high TF in the studied soil. The high TF observed for Cd (4.0 and 1.5 for shoots and fruits, respectively) could be related to the high Ca saturation of the soil because the latter cation can compete in the exchange complex with Cd, enhancing its release to the soil solution and thus plant uptake. This explains the increased TF values for Cd to 5.7 (shoots) and 2.5 (fruits) with PG-inputs in treatment T3. The largest differences in TF for the same element between fruits and shoots (Fig. 2) can be explained by its mobility in phloem. Since fruits have a lower transpiration rate than leaves (included in shoots), the TF must be obviously lower in fruits than in shoots for non-mobile elements in phloem such as Fe [33]. On the other hand, the uptake of Zn is mediated by the nicotianamine aminoacid [34], particularly active in tomato plants [35], promoting the phloem transport and thus the accumulation of Zn in fruit [36,37]. This explains the similar TF for shoots and fruits observed for this element (Fig. 2). A similar mechanism has been reported for Cu [38] which also shows similar TF for both plants tissues. After crop, concentrations of Mn, Mo and Th in soil significantly (at 95% CL) increased with T3 by 9%, 300% and 25%, respectively; meanwhile the concentrations of Ni, Zn, As and Ag decreased by 10%, 19%, 24% and 16%, respectively. The foliar fertilizer did not contribute to elements inventories in substrate, but the Osmocote represents an obvious source of Mo (Table 3), and the added amount can explain the increase of its inventory by a factor four. Changes in the concentration of studied elements after crop can be hardly explained by a mass balance. The amounts introduced with irrigation water were negligible when compared with its inventory present in soil for all analytes, the highest inputs being those of Sb, Mo and Zn (3%, 1.2% and 1.0% of their inventories, respectively). Overall, losses through drainage were much lower than 1% of the Table 6 Transfer factors found in the experiment for control pots versus some bibliographic data: IAEA-TECDOC-1616 [28], PNNL-13,421 [29] and ORNL-5786 [30]. Experiment results IAEA-TECDOC-1616aPNNL-13,421 ORNL-5786 Vegetal tissue Fruit Leafy vegetables Non-leafy vegetables (fruits, berries, . . .) Leafy vegetables Fruit Leafy vegetables Fruits, seeds, tubers, etc.. . . As 0.005 N.M. – – 0.04 0.006 0.04 0.006 Th 0.009 0.018 0.0018 0.0008 0.0018 0.00025 0.00085 0.000085 Fe 0.039 0.0033 0.0010 0.0010 0.05 0.05 0.004 0.001 Al 0.045 0.017 – – – – 0.004 0.00065 V 0.045 N.M. – – – – 0.0055 0.003 Cs 0.049 0.14 0.06 0.021 0.46 0.22 0.08 0.03 Be 0.056 N.M. 0.01 0.0015 0.01 0.0015 U 0.06 N.M. 0.02 0.015 0.0083 0.004 0.0085 0.004 Pb 0.073 0.042 0.08 0.015 0.01 0.01 0.045 0.009 Co 0.078 0.0017 0.17 0.14 0.23 0.007 0.02 0.007 Cr 0.08 0.014 0.0010 0.0010 0.0075 0.0045 0.0075 0.0045 Mn 0.086 0.023 0.41 0.3 0.7 0.05 0.25 0.05 Ba 0.118 0.007 0.005 0.005 0.15 0.015 0.15 0.015 Tl 0.12 N.M. – – 0.004 0.0004 0.004 0.0004 Ni 0.35 N.M. 0.17b– 0.28 0.06 0.06 0.06 Ti 0.37 0.13 – – – – 0.0055 0.003 Sb 0.4 0.72 0.00009 0.00013 0.00013 0.00008 0.2 0.03 Cu 0.43 0.32 – – 0.4 0.25 0.4 0.25 Zn 0.43 0.43 2.4 0.4 1.3 0.9 1.5 0.9 B 0.52 0.33 – – – – 4.0 2.0 Se 1.09 0.28 – – 0.15 0.05 0.025 0.025 Sr 1.1 0.024 0.76 0.36 3.0 0.2 2.5 0.25 Ag 1.13 1.66 0.00018 0.0006 0.00027 0.0008 0.4 0.1 Mo 2.2 4.5 0.5 – 0.8 0.05 0.25 0.06 Cd 3.8 1.5 0.27 – 0.55 0.12 0.55 0.15 aMean geometry (if exist disponibility). bGrass. N.M. results under MDL. – Non reported. final inventory, the highest losses being found for Sb, Sr, Mo and Cd, with 0.8%, 0.6%, 0.5% and 0.4%, respectively. Similarly, extractions by the aerial part of the tomato plants were below 1% of the final inventory in the substrate for all the elements, except for Mo, Cd, Ag and Sb with extractions of 3.8%, 2.8%, 1.7% and 1.1%, respectively. The contribution of plant roots (not measured) and variations in the digestion yield with time remain as possible explanations for these observed changes in the concentrations of studied elements in soil. 4.2. Phosphogypsum effects on analyte concentrations in the soil–plant–water system The added masses of PG in treatments T1, T2and T3represented, respectively, 0.5%, 1.6% and 5.2% of the total mass of the growing substrate. Thus, only the analytes whose concentrations in PG were one order of magnitude higher than those in the soil (B, Cd and U) could appreciably change their effective concentration in the substrate. However, non-significant differences were found with our experimental settings (Table 5, at 95% CL). PG can alter the geochemistry of some elements in the soil, modifying the performance of the microwave-assisted acid digestions. This can explain the differences observed among treatments for the concentrations of Co, Ni, Cu and Pb in soil (Table 5), which were always lower than 9%. The concentrations in shoots of 9 of the 25 studied analytes, namely B, Mn, Co, Ni, Cu, Sr, Ag, Cd and U were significantly affected by PG. Relevant amounts of Cd were supplied with PG, this being the main reason for the observed increase in Cd concentrations in shoots (Fig. 3). On the contrary, PG inhibited the uptake of Mn, Co and Cu, wherein Cu concentration in the shoots monotonically decreases with the increase of PG doses (Fig. 3). This often has been described as a result of antagonism with Ca [39,40], and some studies have suggested the potential use of PG in contaminated soils to reduce the uptake of some heavy metals by plants [41–44]. The lowerPG dose(T1) resultedin decreaseof Uconcentration inshoots, likely as a result of two opposite trends: the inhibition of the uptake by Ca-antagonism, and its increased availability due to its supply with PG. Most of the elements usually show low mobility through phloem, which is the main passage of nutrients from roots to fruits; this explains their overall lower accumulation in fruits in comparison to shoots. Be, V, As and U were under their respective MDL in fruits (Table 1). The highest PG doses (T3) decreased the concentrations of B, Cu, Sb, Cs, Ba, Tl and Th when compared with control. It is worth noting that inhibition of Sb, Cs, Tl, Th and Ba uptake cannot be statistically solved for shoots with the present experimental settings. Decreased concentrations of Mn and Co at increasing PG doses were found in shoots, but not in fruits. The Ca-antagonism can be argued as the main reason for such observed decreased concentration of metals in plant tissues. Iron concentration in fruits tended to increase with increasing PG doses, perhaps due to the supply done with the amendment; on the contrary, Pb concentration tended to decrease with increased PG doses (Table 5), as the likely result of the precipitation of Pb phosphates promoted by P supply with PG amendment [45,46]. As described in shoots, Cd concentrations in tomato fruits increased with increasing PG doses due to its associated inputs (Fig. 3). A similar increase was observed for Se (Fig. 3) which was not significantly supplied by PG amendments. It has been previously reported that P and S (both supplied by PG) can compete for sorption sites with Se in soil [47] and also enhance the transport of Se within the plant system [48,49]. Results of analyte flowing through drainage waters cannot be considered representative of the actual field conditions. Nevertheless, they provide a good basis for a qualitative gradation of the potential risk of loss through leaching of different elements. The highest leachability under our experimental conditions was found for Sb, Sr, Mo and Cd, with cumulative losses through drainage 0 2 4 6 8 10 12 14 0204060 Cu ( mg kg-1 d.w.) Tn-CaPG ha-1 Shoots Fruits 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 020 40 60 Cd ( mg kg-1 d.w.) Tn-C aPG ha-1 Shoots Fruits 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 020 40 60 Se ( mg kg-1 d.w.) Tn-C aPG ha-1 Shoots Fruits 0 0.5 1 1.5 2 2.5 3 3.5 4 020 40 60 Pb ( mg kg-1 d.w.) Tn-C aPG ha-1 Shoots Fruits x 10 Fig. 3. Behavior of concentrations found in fruits and shoots as a function of PG dose (Tn-CaPG ha−1, tons of Ca supplied by PG per ha) for selected elements. Dotted lines have been added to guide the eye. equivalent to 0.8%, 0.6%, 0.5% and 0.4% of their total inventories in substrate, respectively. Overall, the highest PG rate (T3) reduced the cumulative drained volumes by a 40% with respect to the control pots, and thus decreased the associated losses through drainage for most of the elements. This effect can be attributable to an improvement in soil microstructure due to the displacement of Na by Ca in the soil exchange complex. 4.3. Food safety The European regulation 2005/87/CE on undesirable substances in animal feed fixed maximum allowed levels of 30 mg kg−1of Pb in green fodder and 1 mg kg−1of Cd in feed materials of vegetable origin, with a moisture content of 12%. The measured concentrations of Pb in tomato shoots were one order of magnitude below this limit, and values in fruits were even lower. Thus, present data on Pb levels and the potential PG effects on them do not represent any concern. Nevertheless, present levels of Cd in tomato shoots cultivated in these soils (control pots) are below, but close to the maximum allowed concentration, and thus it can be supposed that the application of few additional PG amendments would overcome such a limit. The European regulation 1881/2006 sets maximum levels for certain contaminants in foodstuffs, being those for Pb and Cd in vegetables and fruits of 100 and 50 ␮g kg−1, respectively (wet weight basis). Taking into account the water contents given in Table 4, the levels of Pb were not of concern. Present levels of Cd in tomato fruits cultivated in these soils (control pots) were 38.4 ±2.7 ␮g kg−1 wet weight. Thus, as stated above for animal feed, additional PG amendments would overcome such limit for human consumption, as revealed by results with T2 and T3 treatments, with Cd concentrations in fruits of 64 ±7 and 83 ±15 ␮g kg−1wet weight, respectively. 5. Conclusions The highest soil-to-plant transfer factors (TFs) in tomato were found for Cd, Mo and Ag. Extractions by the aerial part of the tomato plant were over 1% of the final inventory in substrate only for Mo, Cd, Ag and Sb, with values of 3.8%, 2.8%, 1.7% and 1.1%, respectively. The highest leachabilities under our experimental conditions were found for Sb, Sr, Mo and Cd, being always lower than 1% of their total inventories in substrate. Relevant amounts of Cd were supplied with PG (2.1 mg Cd kg−1PG), which explains the observed increase in Cd concentrations in shoots and fruits. The highest PG doses inhibited the accumulation of Mn, Co and Cu in shoots; and of B, Cu, Sb, Cs, Ba, Tl and Th in fruits. The PG effects on Pb concentrations in shoots and fruits do not represent any concern in terms of European regulations for animal feed and human foodstuffs. Nevertheless, present levels of Cd in tomato (shoots and fruits) cultivated in these soils (control pots) were below, but close to the maximum allowed concentrations, and accumulated PG amendments greater than 16 g PG kg−1soil (equivalent to three amendments with 20 Mg ha−1 in the field) would exceed such a limit. The assessment on traceelement content of PG, which may vary depending on the source, is needed to avoid exceeding legal limits of Cd or other pollutants in food when this by-product is used as soil amendment in agriculture. Acknowledgements This work was partly funded by ENRESA (Spanish Public Corporation of Radioactive Residues) and by the IFAPA-C039 project from the Regional Andalusia Government. The authors wish to thank to the Agriculture Cooperative “Las Marismas” for making some facilities and the experimental site available.