| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Consortium for Risk Evaluation with Stake Holder Participation, Environmental and Occupational Health Science Institute, Division of Life Sciences, Rutgers University, 604 Allison Road, Piscataway, NJ 08854
b Consortium for Advanced Radiation Sources, The University of Chicago, 5640 South Ellis Avenue, Chicago, IL 60637
c Savannah River Ecology Laboratory, University of Georgia, Drawer E, Aiken, SC 29802
* Corresponding author (punshon{at}srel.edu)
Received for publication December 2, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Abbreviations: SXRF, synchrotron X-ray fluorescence • XANES, X-ray absorption near edge structure spectroscopy • XAS, X-ray absorption spectroscopy
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The application of spatially resolved microanalysis to plants has generally been confined to the unique metal metabolism of hyperaccumulator plants (Kramer et al., 1997, 2000; Küpper et al., 2001; Salt et al., 1999, 2002), which have evolved the ability to take up very high concentrations of metals (>1% dry weight in the aboveground parts of the plant) such as Ni, Co, Cr, and Mn (Brooks, 1987). Few studies have focused on nonhyperaccumulator plants (Howe et al., 2003), even fewer on field-collected long-lived plants surviving in metal-contaminated areas. This limitation is a result of the high detection limits of SXRF combined with lower metal accumulation characteristics of nonhyperaccumulator plants. However, improvements in beamline technology, such as the use of multiple-detector arrays, has widened the scope of SXRF application and significantly lowered detection limits.
Our understanding of metal transport, distribution, and speciation in nonhyperaccumulator plants, especially long-lived woody plants, is lacking. Baseline information on the spatially resolved concentration of micro- and macronutrient elements in uncontaminated plants is not available, and is becoming increasingly necessary for the validation of data collected from contaminated plants. This information would directly benefit noninvasive, ecologically sensitive remedial efforts such as phytoremediation and short-rotation biomass forestry, as well as aiding in risk assessment and strategic revegetation.
This study follows on directly from an investigation that found extremely high Ni concentrations within an annual ring of a black willow growing on a former radiologically and metal-contaminated settling pond (Punshon et al., 2003a). This individual was resampled, because the concentration of Ni within an annual ring corresponding to 1996 represented a high enough concentration of Ni to study metal speciation within a nonhyperaccumulator plant using X-ray spectroscopic techniques. This individual was growing adjacent to the breached wooden spillway, at the edge of the stream that carried eroded sediments from the former pond—now a successional wetland—into the riparian corridor to the south. The individual potentially came into contact with contaminated sediments as they were exported from the pond, and was surviving in soils with far higher contaminant concentrations than other areas of the former pond (Sowder et al., 2003). Due to the location of the tree, the absence of nearby neighbors of the same species, and the extreme heterogeneity of soil contaminant distribution, it was not possible to obtain comparative samples from other trees.
The existence of two distinct distributional patterns of Ni within rings of black willow was observed using SXRF in the previous study (Punshon et al., 2003a), although the resolution of the analysis was not high enough to discern the nature of the distributions. Areas of diffusely Ni-enriched woody tissue contained localized areas of intense Ni enrichment that measured 10 to 20 µm in diameter. Closer investigation indicated that these features were an exclusively Ni-containing substance within the lumen of xylem vessels. It was suggested that the contrasting distributions represented Ni fixed to the cell walls (diffuse) as well as transported within the vessels (concentrated), and in this paper we hypothesize that they may differ in their Ni speciation characteristics.
Here we apply SXRF microanalysis and XAS to metal-enriched woody tissue collected from a tree known to be impacted by soil metal pollution (Pickett, 1990; Punshon et al., 2003a) and investigate the distribution, co-association, and chemical binding environment of Ni within enriched annual rings.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Sample Collection and Preparation
Core samples were collected in early 2003 from an individual known to have been impacted by Ni from previous investigations (Punshon et al., 2003a, 2003c). Previous one-dimensional SXRF line scans showed elevated Ni in the 1996 annual ring only, therefore this ring was the focus of the study. Woody stem tissue was collected from an 8-yr-old black willow tree, growing adjacent to the wooden spillway at Steed Pond. Five replicate core samples were collected 90 cm above the base using a Teflon-coated Mattson increment corer (20-cm length, 0.5-cm i.d.) (Ben Meadows, Janesville, WI). Samples were placed in resealable polythene sample bags and transported back to the laboratory, making careful note of orientation, so that the age of annual rings could be determined.
One sample was glued to a grooved pine block and polished with progressively finer grades of sandpaper until rings were easily distinguishable. The rings were measured with a Henson University model incremental measuring machine (Fred C. Henson Co., Mission Viejo, CA) with attached stereo microscope. The locations of annual ring boundaries were noted in millimeters.
For SXRF analysis, intact cores were cut into three 25-mm sections with a stainless steel surgical blade and embedded in 3M (St. Paul, MN) Scotchcast electrical resin (trace-metal-free epoxy liquid resin) and core sections were positioned to maintain correct orientation with respect to pith and cambium, and so that wood fibers (i.e., vessel elements) were perpendicular for all three sections. Sections were sliced transversely with a diamond wafering saw (Buehler, Lake Bluff, IL) to provide samples 1 mm thick. For SXRF analysis the surfaces of the sliced samples were polished with 600-grit special silica carbide grinding paper for metallography (Buehler) before a final polishing stage using Metadi supreme polycrystalline 3-µm water-based diamond suspension (Buehler) to remove any superficial metals. Only core samples intended for age determination came in to contact with conventional sandpaper.
Synchrotron X-Ray Fluorescence Analysis
The SXRF microbeam analysis was conducted at the X26A beamline at the National Synchrotron Light Source, Brookhaven National Laboratory (Bertsch et al., 1994), and was used in the compositional (two-dimensional) mapping mode for Ni, Cu, Zn, Mn, and Fe. The incident X-ray beam was tuned to 17.5 keV using a Si (111) channel-cut monochromator. During the period of analysis the beam was focused to 10 x 10 or 5 x 10 µm using rhodium-coated Kirkpatrick-Baez focusing optics (Eng et al., 1998). Energy dispersive XRF compositional data were collected using a Canberra Industries (Meriden, CT) SL30165 Si (Li) detector.
Quantification of elemental abundances (mg kg–1) from fluorescence counts was based on analytical procedures developed at X26A (Lanzirotti and Miller, 2002), and refined from earlier studies (Punshon et al., 2003a). A sample of black willow leaf material was prepared for standardization purposes. The leaf material was dried (60°C), homogenized in a cryo-mill (Spex; CertiPrep, Metuchen, NJ), and vacuum-compressed (15 x 105 Pa for 5 min) into pellets. Elemental composition of microwave digested material (MDS 2000; CEM, Matthews, NC) was determined using an Elan 6100 DRC Plus (Perkin-Elmer, Wellesley, MA) inductively coupled plasma mass spectrometer (ICP–MS). Standard reference materials SRM1515 (apple leaves) (National Institute of Standards and Technology, Gaithersburg, MD) were included in ICP–MS analysis, and demonstrated recoveries between 75 and 105%.
A standards-based, fundamental parameters approach was then used to calculate elemental abundance in the tree cores based on the measured fluorescence yield. This was done using a version of the public-domain software NRLXRF (Criss, 1977) specifically modified for synchrotron X-ray fluorescence. Corrections were made for differences in sample density and thickness, absorption of the incident beam by Be windows, photoionization efficiencies, fluorescence yields, self-absorption, secondary fluorescence, and fluorescence beam absorption by air and detection filters (Sutton et al., 2002). Comparison of calculated abundances for the black willow leaf standard with ICP–MS data yielded good agreement. In using this approach beamline optical parameters and sample thickness are very well constrained. However, changes in sample density along the length of the tree core will yield differences in fluorescence yield that can potentially be misinterpreted as changes in elemental abundance. For example, a 1-mm-thick section of willow with a composition similar to cellulose would yield an increase in fluorescence intensity of roughly 1.5 times if the wood density increased by 2.5 times, from 0.4 to 1.0 g cm–3; a typical density fluctuation (Schweingruber, 1988). However, it is possible to directly constrain the variability in material density (
in g cm–3) by measuring the transmitted (I) and incident (I0) X-ray intensities since they are a function of the mass absorption coefficient of the sample (µ, approximately 3.9 cm2 g–1 for cellulose) and its thickness (d) through the relationship:
 |
This has been performed in Fig. 1, as a line scan across the core section used throughout this study, perpendicular to the annual rings. These measurements yield a maximum calculated density of 0.52 g cm–3 and a minimum calculated density of 0.42 g cm–3, a 1.2x difference. The measured fluorescence intensity for Ni K
, however, varies by 4.5x and the Fe K intensity by 188x, variations too large to be accounted for by changes in sample density alone. Thus, for abundance calculations we believe assuming the sample has a uniform average density of 0.42 g cm–3 introduces only minimal uncertainty. Detection limits for most of the elements analyzed were within the range 0.1 and 10 mg kg–1. Compositional maps were produced using a step size of 30 µm and a dwell time of 3 s pixel–1 (unless otherwise specified).
|
The energy step sizes used varied, with 0.1-eV steps used in the pre-edge and main edge region to extract good-quality XANES spectra, and 2- to 5-eV steps after the main edge region. For standards, collection times varied between 2 to 4 s per point. For the tree samples collection times were up to 30 s per point and 10 scans were averaged together to improve the signal to noise ratio. Energy calibrations for the position of the Mn and Ni K absorption edges were made based on repeated analyses of Mn and Ni metal foils. Spectra were analyzed using the IFEFFIT program (Newville, 2001) and the ATHENA interactive graphical utility (Ravel and Newville, 2005) for XAFS data processing.
Cluster Recognition Analysis
We used spatially explicit analyses to gain an understanding of elemental co-associations in contrasting regions of the woody stem tissue sample. Object-based classification of spatially resolved metal abundance data is far more informative than sample-wide analyses, because it allows clusters (i.e., contiguous areas with relatively homogeneous counts) to be delineated in a standardized manner, and removes potentially biased, subjective interpretation (Vogt et al., 2003). Groups, rather than individual pixels, are classified as objects, allowing both spectral and contextual information to be incorporated while reducing inherent pixel-based noise. Once data images are segmented into unique clusters, further comparisons can be made both within and between clusters. We used matrix filters and advanced learning algorithms incorporated in Feature Analyst (Version 3.4; Visual Learning Systems, 2001), an extension of ArcGIS (Version 8.3; ESRI, 2002), to conduct unsupervised classifications of elemental composition maps. Clusters were identified for Ni, the primary contaminant of concern, and then distinct regions of high and low counts were overlaid on maps of the remaining elements.
Statistical Analysis
Statistical analyses were performed both on individual pixel data for the entire sample (i.e., not clustered) and pixels extracted from the entire sample on the basis of their similarity (clustered). For nonclustered data, descriptive statistics and correlations between metals were ascertained. For clustered data, differences between distinct clusters were determined using general linear models, and within-cluster variation was quantified using a coefficient of variation measure (CV).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
Elemental abundances of Ni were within the range 0 to 77 mg kg–1, with the highest concentrations located within the clearly delineated boundary of an annual ring between 0 and 8 mm (Fig. 2B). According to dendrochronological measurements, this ring was formed in 1996, and corresponds to the enriched annual ring from the same individual from the previous study (Punshon et al., 2003a). Over the entire core section, median Ni concentrations were 12.9 mg kg–1. Elevated Ni concentrations were found in smaller localized areas distributed throughout the sample, and were marked for more detailed investigation. Nickel was present at a low concentration at the 1995–1994 ring boundary at 20 mm, relative to the surrounding tissue. Of all elements of concern, Ni and Mn were most consistently enriched over a comparatively greater area of the sample.
The sample-wide median value for Mn was 177 mg kg–1 (Fig. 1E). Adequate bulk tissue concentrations of Mn in plant tissues is approximately 50 mg kg–1, and although there are problems associated with comparing volume-averaged and spatially resolved metal concentrations, the concentrations observed here appear elevated, consistent with the elevated Mn concentrations found in Steed Pond soils. Spatially resolved pairwise correlations conducted among all elements of concern showed that the most significant correlation was between Ni and Mn, where r = 0.8822.
Median Cu and Zn concentrations were low throughout the core section, consistent with their levels in Steed Pond sediments. For both elements, the higher concentrations were located in the 1996 annual ring (0–8 mm) and (unlike Ni) in the boundary between the 1995 and 1994 annual rings. The majority of Cu observed fell within normal tissue concentrations (6 mg kg–1 dry weight is considered adequate) (Salisbury and Ross, 1992) and the sample-wide median Cu value was 3.6 mg kg–1 (Fig. 2C).
Maximum Zn concentrations reached 1038 mg kg–1 in isolated regions, with a sample-wide median of 23.5 mg kg–1 (Fig. 2D). Tissue Zn concentrations of 20 mg kg–1 are considered normal (Salisbury and Ross, 1992). Although small isolated areas of Cu and Zn existed within the sample, they did not correspond spatially with one another, or with Ni, and are disregarded as outliers in statistical analyses. Additionally, the sample-wide correlation between Cu and Zn was weak (r = 0.119).
Adequate volume-averaged tissue Fe concentrations in plant tissues are in the range of 100 mg kg–1 (Salisbury and Ross, 1992). Concentrations of Fe were elevated in specific regions of the sample (Fig. 2F), with a maximum value of 10821 mg kg–1. Concentrations of Fe in Steed Pond sediments were also highly elevated above that of the control site. Consistent with observations from Fig. 2, sample-wide correlation showed that Fe was negatively correlated with Cu, Mn, and Ni, with r values of –0.21, –0.22, and –0.38, respectively. Iron was, however, positively correlated with Zn, with r = 0.52.
Cluster Recognition Analysis
Only Ni, Mn, and Fe were considered in CRA analyses because Cu and Zn were not present at elevated concentrations. The mean concentration of Ni in the 1996 annual ring was 26.1 (±SD 7.1) mg kg–1 (median = 26.0 mg kg–1), whereas in 1995 this was 13.7 (±4.1) mg kg–1, a significant difference (P = 0.0001, F1 = 49294). In 1996, the correlation between Ni and Mn was r = 0.42, whereas in 1995 this correlation was weaker (r = 0.12).
Mean Mn concentration in 1996 was 281 (±57) mg kg–1 compared with 153 (±38) mg kg–1 in 1995 (P < 0.0001, F1 = 73320). Coefficients of variation for Ni were 27 and 30% for 1996 and 1995, respectively, whereas for Mn the coefficient of variation is 20% in 1996 and 24% for 1995.
Due to the contrast between Fe and the other elements of concern, Fe data were analyzed using a separate CRA, based on the unique clusters delineated for Fe, which bisected the sample in half, with 0 to 13 mm having the lower Fe concentrations, and the remainder of the core elevated in Fe. This gave means of 806 (±710) mg kg–1 and 13151 (±9508) mg kg–1, respectively. Coefficients of variation for Fe data remained elevated; 72 and 88%, respectively.
X-Ray Absorption Spectroscopy Analysis
The XANES data presented in this study were collected from an area of elevated Mn and Ni abundance found in a xylem vessel of black willow. Direct visual comparisons of detailed regions of XANES spectra were conducted using ATHENA interactive graphical utility software (Ravel and Newville, 2005). Although Ni is the primary contaminant metal of concern, we evaluated both Mn and Ni because the high correlation between these two elements in the enriched annual ring suggests either that they share similar binding sites or are similarly speciated. Manganese and Ni can occur in a wide range of valencies. The most stable states for Mn are typically 2+, 4+, and 7+, but some insoluble 3+ compounds can be also stable, as can some 5+ and 6+ states in strongly alkaline solutions. The most common oxidation state of Ni is 2+, although 1+ through 4+ Ni complexes do exist. While Ni2+ is the most common valence in nature, it can be found in a wide variety of coordination geometries. The energy position of the main edge in these XANES spectra is a function of the length of the Mn or Ni to anion bond and is thus sensitive to the oxidation state of the absorbing atom. Figures 3 and 4 show normalized Mn and Ni K-edge XANES spectra, respectively, of several Mn and Ni compounds in comparison with XANES spectra collected from black willow. Spectra were also collected from woody tissue from both the 1995 and 1996 annual ring and, other than the lower intensity, no appreciable differences in spectral shapes were noted. For both Mn and Ni compounds the energy of the main absorption edge shifts to higher energy with increased valency. The edge positions for Mn and Ni measured in this annual ring of black willow are both similar to those measured in other 2+ species.
|
|
3d, and are also sensitive to changes in the coordination geometry of the absorber. Typically, the higher the symmetry, the weaker the observed pre-edge structure. Since the height and position of these peaks in Mn and Ni compounds is a reflection (at least in part) of the coordination of the absorber (Farges et al., 2001; Feth et al., 2003; Galoisy and Calas, 1993; Nietubyc et al., 2001; Wong et al., 1984), comparisons of pre-edge energy and normalized peak height allows for distinction between various coordination geometries (Fig. 4A). Such modifications in XANES spectra are distinctive enough to allow for fingerprinting of the metal species by comparison with spectra from likely standard compounds. The XANES spectra collected from Ni within black willow woody tissue were similar to several Ni-bearing standards (Fig. 4B). When the spectra were overlaid on one another using imaging software, the pre-edge peak and remaining spectra of a prepared Ni–pectic acid complex were most similar to that observed in black willow. However, detailed comparisons between the Ni–histidine and hydrated sulfate spectra and the compound found within black willow indicated that these forms of Ni could not be completely ruled out. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Previous studies on the partitioning and mineralogy of Ni in Steed Pond soils have shown that Ni and Mn are strongly spatially co-associated (Sowder et al., 2003). It is thought that both Ni and Mn substitute for Fe on the iron-oxide structures that form a predominant component of the soil mineralogy in this area, because their cations have very similar crystalline radii (Ni2+ = 0.069 nm, Mn2+ = 0.060 nm) to that of Fe3+ (0.064 nm). Since Steed Pond is a seasonal wetland, fluctuations in redox conditions encourage dissolution and reprecipitation of iron oxides, resulting in iron oxides that are highly substituted with suitable metals. Although these metals are clearly not incorporated within the tree as oxide species, we can speculate that their close association may have carried over into the plant tissue, and the similarity of their ionic radii may have resulted in their being bound to similar ligands during transport within the vascular system.
Manceau et al. (2002) also reported a close association of Ni and Mn in soil ferromanganese micronodules collected from a French agricultural soil. The formation of these micronodules is an important metal sequestration process in many soils, and Mn and Ni were systematically co-associated in all of the nodules examined. Varga et al. (1999) note that Ni contamination hinders the transport of K and Zn, and leads to accumulation of Mn in root tissue. In the Ni hyperaccumulating species yellow tuft (Alyssum murale Waldst. & Kit.), Broadhurst et al. (2004) also report a frequent co-association between Ni and Mn.
From the XANES spectra collected during this study, the species of Ni present in the vascular system of black willow may exist either as a Ni–pectic acid complex or a Ni–histidine complex. Comparison of the spectra also shows less pronounced similarities with the hydrated sulfate form of Ni. Conceivably the form of Ni within black willow may not be represented by any of the standards used in this study, the selection of which was based on previous studies of Ni speciation conducted on hyperaccumulators (Kramer et al., 2000; Salt et al., 2002), and from an understanding of vascular system chemistry.
A Mn–pectic acid complex and Mn–histidine complex were not included in the suite of Mn-bearing standards, and therefore a conclusion about Mn speciation cannot be made. The data presented here confirm a strong co-association between Mn and Ni, which may be the result of these elements being similarly speciated, or more simply it may imply that the availability of Mn and Ni over time is similar.
Conversely, the similarity between XANES spectra of the Ni–pectic acid complex, the Ni–histidine, and the Ni within black willow xylem vessels may arise from a mixture of Ni species within black willow. Pectic substances are straight chain aliphatic polymers that are present in various structures within the woody vascular system, as cementing agents between the cells walls of the xylem elements, xylem ray cells, and the end walls of vessels and tracheids (Momoshima and Bondietti, 1990), conferring a high metal binding capacity to these structures. Histidine is thought to play an important role in Ni transport, for example, in the hyperaccumulator Alyssum sp. (Kramer et al., 1996). Nickel moves as a complex within the transport system, although free ions may be released and bind to fixed, negatively charged carboxyl groups lining the lumen of the xylem vessels. The Ni within black willow could conceivably contain both forms, and may be Ni–histidine adhering to an end wall of the xylem vessel as a result of removal and drying of the wood sample. The Ni–pectate complexes may represent smaller fractions of Ni that move between the cell walls of the xylem elements, whereas the majority of Ni present in the xylem sap is released from histidine (or potentially another ligand) to bind to fixed sites lining the xylem vessels.
Data on the Fe concentration range or distribution normally found within the vascular system of woody plants, specifically willows, is sparse in the scientific literature, although physiological texts suggest a plant-wide average of 100 mg kg–1 (Salisbury and Ross, 1992). Data from previous SXRF analyses and ecological studies performed at Steed Pond (Punshon et al., 2003a) show that Steed Pond soil contains far more Fe than the control site (Punshon et al., 2003b, 2003c). The concentration of Fe within annual rings of control-site willows was within the range of 0.07 to 14.74 mg kg–1, with an average concentration of 2.2 (±SD 2.3) mg kg–1, whereas in this study, a sample-wide median Fe concentration of 59 mg kg–1 was punctuated with a band at the 1995–1994 ring boundary where the concentration reached several thousands of mg kg–1, concurrent with a lack of Ni and Mn. There are several potential causes for the observed variability in the Fe concentration and its contrasting distribution to Ni. First, periodic flooding of wetland soils of Steed Pond causes dissolution and reprecipitation of Fe oxides, influencing Fe availability, whereas Ni is not a redox-sensitive metal. Second, the role of Fe and Ni within plants differs. Iron is a plant essential nutrient required in higher concentrations than Ni, Cu, and Zn. In a previous study, comparison of Fe distribution in tissues with photomicrograph images indicated it was a widely distributed structural element (Punshon et al., 2003a), and in Fig. 2 the striated distribution of Fe at 20 mm clearly shows the path of vessel elements. Nickel, however, is an essential micronutrient, and is a constituent of urease enzymes, which are involved in nitrogen metabolism (Gerendas et al., 1998).
Finally, this study suggests that SXRF can reveal important distributional characteristics of metals within nonhyperaccumulator plant tissues, and in this instance, a "record" of Ni appeared to have been conserved within an enriched annual ring. However, interpretations with respect to retrospective biomonitoring must remain tentative. Analysis of excised, air-dried tissue presents some uncertainties for XAS analysis. The in vivo woody vascular system is a continuous column of solution, which is interrupted during sampling and further affected during drying. It is likely that the introduction of air into the vascular system breaks the column, causing solutes to contract into what is observed as small inclusions when analyzed with SXRF. Further, areas with high metal binding capacity, such as the pectic acid–enriched end walls of vessel elements, may be areas where metals are most likely to adhere during drying, or where metals are most likely to be found at sufficient concentrations to be detected using SXRF. The solution to this issue is to analyze plants in vivo, which has recently only proven possible with small hyperaccumulator plants (Scheckel et al., 2004), although combined with computed microtomography, artifacts arising from sample collection and preparation can be eliminated.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Related articles in JEQ:
This article has been cited by other articles:
|
|
 |
|
  S. A. Kim, T. Punshon, A. Lanzirotti, L. Li, J. M. Alonso, J. R. Ecker, J. Kaplan, and M. L. Guerinot Localization of Iron in Arabidopsis Seed Requires the Vacuolar Membrane Transporter VIT1 Science, November 24, 2006; 314(5803): 1295 - 1298. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Vadose Zone Journal | Journal of Plant Registrations | ||||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal |
