Chrysanthem u m cinerariaefoliu m is a pere nnial plant, herbaceous, multishoote d, and belonging to Asteracea family. This plant is the principal source of  pyret hrins, the most eco nomically im portant natural insecticides (16). They offer all of the qua lities of ideal pest control agents (i.e., rapidity of action, activity against a broa d range of insects, and low cost) and, beca use of their rapid biodegradab ility, some of the advantages of biological agents too, such as weak development of resistant str ains and low toxicity for mamm als (17). It is an im portant comm ercial plant; the world market size is estimated to be as high as $400 million a year; the wholesale price of flower extracts , on a basis of 100% pyret hrin conte nt, was reporte d to be $413 / kg. The global market volu me is more than 12 000 to ns of dry flowers / year, while the world demand for dry flowers re mains at 20 000 to ns/ year. Conventional produ ction of pyret hrins still falls short of world require ments, despite an incre ased produ ction of Chrysanthem u m cinerariaefoliu m flowers and pyret hrins, especially in Austr alia (16).

Materials andMethods

Laboratory Pilot Plants. The labo ratory pilot plants consisted of poly(vinyl chloride) ta nks 4 m long, 0.15 m wide, and 0.10 m deep. The pu rification system used an NFT soilless culture

(15) with the per manent recirculation of 30 L of wastewater, regulated by an electric pu m p with a flow rate of 10 L min-1 (Figure 1).

10.1021/es011323d CCC: $22.00 2002 American Chemical  Society VOL. 36, NO. 9, 2002 / ENVIRONMENTAL SCIENCE &  TECHNOLOGY 9     2101

Published on  Web 04/05/2002

FIGURE1. Systemof wastewatertreatment: horizontal flow system withplants (Chrysanthemum cinerariaefolium) in the channel (P is the pump).

effective length) with a 75 m m inter nal diameter. All measure ments were made in anionic mode (19). The capillary was ther mostated at 25 C, and a constant voltage of 20 kV, with an initial ram p of 0.17 min, was applied du ring analysis. Sam ple injections were  made with press ure mode for 30  s,

0.5  psi. The indirect  UV detect ion was at  254  n m  with  a ba ndwidth of 1 n m. The carrier bu ffer was a chromate electro lyte solution, 4.7 m M sodiu m chromate (Fisher Scientific, Pitts bu rgh, PA), 4 m M OFM-OH (Waters, Milford, MA), 10 m M CHES (Sigma, St. Louis, MO), 0.1 m M calciu m gluconate (Sigma) at  p H 9.

Kinetics of SS, BOD5, COD, TP, and TN (March, April, and June 2000 ). The re moval of biodegradab le constituents such as SS, BOD5, COD, TP, and TN in wetland systems can be described by the following first- order kinetic   model:Chrysanthem u m cinerariaefoliu m plants were developed

initially in inpidua l pots containing a 50 / 50 mixture  of ver miculite / com post. After 6 months of culture, they were tra nsferre d to a hydroponic system. They have been cultivated in 1 month with a nu trient solution of Lessaint and Co¨ıc (18). The plants grew bare-r oote d in 3 m m solution flowing by gravity. These plants prese nted an average shoot length of 20 cm. Three chann els were used: one with 25 plants supp lied with wastewater (planted chann el), one with 25 plants supp lied with nu tritive solution of Co¨ıc-Lessain t (control plant chann el), and one plant-free chann el supp lied with wastewater (unp lanted chann el) to know the role of plants. These experiments were performed in a glass house with a controlled te m perat ure at 25 ( 3C/ 15 ( 3C (day/ night) and with a  natural ph oto period.

The raw efflue nt was a combined stormwater/ sewage wastewater, coming from a comm unity of 980 people without indu stry; it was represe ntative of a rural comm unity. The raw efflue nt was collecte d weekly and stored at 4 C. Planted and unp lanted chann els were supp lied with 30 L of collecte d raw efflue nt, re newed every 72 h. The experiment lasted 6 months (February to July); this period includ es three seasons and thus differe nt qua lities of wastewater. Wastewater qua lity was followed once a month du ring a 72 h recirculation cycle. In March, April, and Jun e (weeks 5, 10, and 20, res pectively), wastewater sam ples were analyzed after 0, 6, 12, 24, 48, and 72 h of treat ment. In February, May, and July (weeks 1, 15, and 25, res pectively), wastewater samples were analyzed after 0  and 48  h  of treat ment.