Methods were previously described in [29] . The plants used in this study were iris (Iris savannarum), switchgrass (Panicum virgatum), Tithonia rotundiflora, Coreopsis lanceolata, sunflower (Helianthus annuus), and marigold (Tagetes erecta). A 25% perlite, 37% pine bark, 8% sand, 30% coir potting mixture was used for all plants except iris. Ten cm iris rhizomes, collected from a single plant were set in rockwool to help maintain rhizome orientation during ebb and flow cycles in the hydroponic system. Switchgrass seed was evenly sowed in 28 × 53 cm trays. Once plants reached 10 cm in height, 30, 18-cm sections of turf were cutout and placed into 26-cm diameter pots (3.8 L). Switchgrass was trimmed to a uniform 15 cm height before treatments began. Tithonia, sunflower and marigold seedlings with at least two fully developed leaves, and iris and coreopsis plants,10 cm tall with ≥3 leaves, were placed in 26-cm pots. The study was conducted over three different time periods with two plant species growing during each period. Dates for each time period are given in Table 1.

Six ebb-and-flow type hydroponic plant maintenance systems were used for the study. Each system contained a 208 L reservoir tank filled with 132 L water. Each tank was connected to 12, 3.8-L pots. A timer allowed the system to cycle between 30 min. wet and 4 hr. drain periods, beginning at 8 A.M., ending at 4 P.M. followed by a 12 hr. drain period. A modified Hoagland solution was used to supply plant nutrients. Nutrients were added in the form of concentrated stock solutions before tanks were brought to their final volume. Final nutrient concentrations in each tank were 2.0 mM Ca(NO₃)₂, 3 mM KNO₃, 1.0 mM MgSO₄, 0.25 mM Ca(H₂PO₄)₂, 12.5 uM H₃BO₃, 1.0 uM MnSO₄, 1.0 uM ZnSO₄, 0.25 uM CuSO₄, 0.2 uM (NH₄)₆Mo₇O₂₄, and 10 uM Fe-EDDHA. Tap water used to mix nutrient solutions averaged 0.0028 mg∙L⁻¹ As (0.448 mg per reservoir tank). Plants were acclimated to hydroponic feeding for a minimum of one week before beginning As treatments.

An As solution concentration of 2 - 8 uM equates to a soil AC of 700 - 3000 mg∙kg⁻¹ [30] . Based on this and the levels of contaminated urban soils reported above, a range of 10 - 70 uM As solution concentration was selected to cover the range of low level As contamination found in south Florida’s urban soils. Enough Na₂HAsO₄, dissolved in 1.0 L water, was added to different reservoirs to make a final tank concentration of 0, 10, 20, 30, 40, 50 or 70 uM As (0.0, 0.75, 1.5, 3.0, 3.75, 5.25 mg∙L⁻¹ As, respectively). Reservoir pH was adjusted daily to pH 6.5 with either NaOH or H₂SO₄. Nutrient and As solutions were replaced weekly. Plants were maintained in hydroponic solution until flowering.

Shoot and root tissue were harvested separately. Roots were washed with a gentle spray to remove debris, agitated in a pool of water then washed a second time. Shoot and root tissue were oven dried at 45˚C until there was no longer a weight change with additional drying and the dry weights recorded. Dried tissue was stored for analysis. Approximately 0.25 g of oven dried plant tissue was placed in 100-mL digestion tubes. Ten mL HNO₃ was added and samples digested in a microwave digestion system for 15 min to reach 200˚C then kept at this temperature for an additional 15 min. Digests were diluted to 100 mL and stored at 4˚C prior to analysis. Element concentrations were determined by inductively coupled plasma-optical emission spectrometry with an iCAP 6300 Duo View (ThermoFisher Scientific, West Palm Beach, Florida). Data were analyzed and concentrations determined using ThermoFisher Scientific iCAP 6300 iTEVA software. A translocation factor (TF) was calculated as:

TF = shoot As/root As in mg As kg⁻¹ plant dry weight (1)

for each species and each treatment.

Each plant species was analyzed separately. The data represent means calculated from six replicated pots for

each As treatment. Analysis of variance was performed using the Proc Mixed procedure of Statistical Analysis System [31] . Tukey adjusted least square means were used for comparison at P < 0.05 unless stated otherwise. Arithmetic means were used to calculate translocation factors.

Both coreopsis root and shoot dry weights declined with increasing solution AC (Figure 1(a)). Coreopsis 0.0 As control treatment produced at mean dry weight of 6.9 g with a shoot-to-root ratio of 2.52. Healthy roots could support 2.5x their weight in above ground dry matter. At 0.75 mg As L⁻¹ solution there was a 32% reduction in dry weight; reductions for 2.25, and 5.25 mg As L⁻¹ were 65% and 84%, respectively. A significant drop in dry weight began at 2.25 mg As L⁻¹; however, up to that point the allocation of dry matter between shoot and root tissue (shoot-to-root ratio) remained in a range from 2.5 and 2.3. As plants became smaller the loss in dry weight was equally shared by shoot and root tissue. A concentration of 5.25 mg As L⁻¹ resulted in a drop in shoot-to root ration to 1.5. At this level of exposure, roots damaged by As had dropped below a critical level and a proportional shoot weight similar to that produced in the control could no longer be maintained.

Tithonia produced a mean dry weight in the control of 34.2 g (Figure 1(b)). Addition of as little as 0.75 mg As L⁻¹ resulted in a 79% reduction in dry weight. In contrast to coreopsis, the dry weight shoot-to-root ratio for tithonia change very little with an increase in solution As, ranging from 1.11 to 0.98 decreasing steadily from 0.0 to 3.75 mg As L⁻¹, respectively.

In iris, a solution AC of 0.75 mg∙L⁻¹ reduced total plant dry weight accumulation to 50% of the control (Figure 1(c)). However, higher ACs up to 5.25 mg∙L⁻¹, increased plant dry weight. This was true for both root and shoot tissue. Whole plant dry weight increased over controls by 1.8% and 2.2% with solution ACs of 3.0 and 5.25 mg as L⁻¹, respectively. The initial decrease in dry weight at low AC and subsequent increase in dry weight at higher ACs may result from complex interactions with other elements (Fe and Mn) during uptake and translocation. These interactions at low AC might inhibit growth but all defensive mechanisms may not be fully mobilized until solution AC increases. This will be discussed in more detail below. In addition to the reduction in dry weight, the shoot-to-root ratio slightly decreased from 3.0 in the 0.0 As control to 2.5 in the 0.75 mg As L⁻¹ solution treatment. As dry weight increased with higher solution AC, shoot-to-root ratio returned to levels near those found in control plants, 2.8 and 3.0 for the 3.0 and 5.25 mg As L⁻¹ solution treatments, respectively.

Marigold’s dry weight was statistically similar in all treatments for both shoot and root tissue (Figure 1(d)). Shoot-to-root ratios were higher in As treatments than that in the control. Sunflower performed similar to marigold in that there were no differences in dry weight between treatments (Figure 1(e)). Sunflower dry weights tended to decline with increasing solution AC. Shoot-to-root ratios were higher in plant treated with 0.75 and 3.75 mg As L⁻¹ solution As. Switchgrass dry weight decreased with increasing solution AC (Figure 1(f)). The drop in shoot dry weight was greater than that in roots resulting in a constant drop in shoot-to-root ratio with an increase in solution AC.

Arsenic accumulated by coreopsis plants tended to remain in root tissue (Figure 2(a)). As solution As increased, the shoot-to-root ratio of As accumulation went down. Maximum As uptake occurred at 2.25 mg As L⁻¹ solution; above that level shoot As began to decline faster than root As. Switchgrass also accumulated As mostly in root tissue and the shoot-to-root ratio decreased with increasing solution As (Figure 2(f)). Very little As was taken up by iris plants (Figure 2(c)). More of the As taken up by iris was translocated to the shoot in all but the 3.0 mg As L⁻¹ solution treatment. A high sensitivity in tithonia to As, as seen in the reduction in dry weight at 0.75 mg∙L⁻¹ solution concentration resulted in low As accumulation in plant tissue (Figure 2(b)). The shoot-to-root ratio of As accumulation increased from 0.3 mg in controls to 0.9 and 1.0 in the 0.75 and 2.2 mg As L⁻¹ solution treatments. At 3.75 mg∙L⁻¹ plant uptake dropped to 0.4 mg As. Although there were no statistical differences in marigold root As, uptake tended to increase with increasing solution concentration. Marigold shoot As was higher than that in controls, however, less As was partitioned to shoot tissue with increasing solution concentration (Figure 2(d)). Sunflower root and shoot As content increased with increasing solution AC (Figure 2(e)). The shoot-to-root ratio for As was 0.9, 0.7, 0.8 and 0.8 in the 0.0, 0.75, 3.75 and 5.25 mg∙L⁻¹ treatments, respectively.

The AC in coreopsis root and shoot tissue increased with increasing hydroponic solution AC (Figure 3(a)). Considering the continuous decrease in dry weight combined with an increase in As uptake to a solution concentration of 2.25 mg As L⁻¹, coreopsis plant protective mechanism against As were likely overwhelmed at this point. A higher AC was found in coreopsis root than shoot tissue. Maximum tithonia tissue AC was reached at 2.25 mg As L⁻¹ (Figure 3(b)). Tithonia maintained a high shoot AC in all treatments.

Iris whole plant AC increased at 0.75 mg∙L⁻¹ solution concentration due to an initial drop in plant dry weight (Figure 3(c)). Subsequent increases in tissue AC were due to increased As uptake. Shoot AC increased with increasing solution As. At 5.25 mg As L⁻¹ solution plant uptake was highest; however, AC declined due to a relatively small increase in root dry weight accompanied by greater transport of As to shoot tissue.

Marigold whole plant tissue AC peaked at 3.75 mg As L⁻¹ solution (Figure 3(d)). At a hydroponic solution concentration of 5.25 mg As L⁻¹, shoot tissue AC dropped off due to a lower dry weight produced by that treatment. Switchgrass had a similar peak at 3.0 mg As L⁻¹; however, above this level a decrease in both As uptake and dry weight reduced tissue AC (Figure 3(f)). Sunflower AC peaked at 3.75 mg∙L⁻¹ solution As; however, there was a slight increase in shoot AC with an increase in solution As to 5.25 mg∙L⁻¹ (Figure 3(e)).

In coreopsis all elements trended to a decrease in root content with an increase in solution As (Table 2). Differences between zero As controls and 5.25 mg As L⁻¹ solution were significant for all elements except Fe and Mo. For Ca, Cu, K and P significant differences appeared at 0.75 mg∙L⁻¹ solution As. A similar decrease in shoot element content with increasing solution As was found in shoot tissue. Shoot tissue had higher element content than root tissue for all elements except Fe and Mo (Table 2).