nutrient solution   wastewater

plant (g dry weight/plant) 35 ( 6 37 ( 7

shoot dry weight/root dry weight 3.2 ( 0.9 3.4 ( 0.7 total pyrethrin s (mg g-1 dry weight)   1.20 ( 0.01 1.16 ( 0.07 pyrethrin s I/pyrethrin s II 1.90 ( 0.03 1.58 ( 0.35

a Values are means ( standard deviation (n ) 6).

FIGURE6. Chlorophyll a fluorescence parameters: (A) efficiency of ΦPSII and (B) ratio of variable to maximum fluorescence Fv/Fm measuredonleafofChrysanthemumcineriaefoliumcultivated with wastewater(b) and witha nutrient solution(O). Values aremeans ( standard deviation; n ) 20.

pyret hrin conte nt in capitula of the seedlings cultivated on wastewater and nu tritive solution showed that these two values were identical (Table 3). Additionally, the pyret hrins

I / pyret hrins II ratio showed no differe nce betwee n the two

growth  types.  As  com pared  with  field  grown  plant,  we

in the NFT system used some of the water, which would have res ulted in nu trient being more conce ntrate d. Kadlec has shown that the evapotranspiration may have significant effect s and influe nce treat ment efficie ncy (38). Thu s, we can assu me that the mass of TN and TP re moved were abo ut 30% higher than that reflecte d by obtained conce ntrat ion.

A principal component analysis (PCA) of these parameter s is represe nted in Figure 5. The first two principal components (axes 1 and 2) accoun t for 86% of inertia. Axis 1 represe nts most of the variance of the res ult (65%) and group s most of the parameter s character istic of the organic bu rden. All of the variables are str ong and positively correlated (decrea se with treat ment time) except for O2 and NO3-, which were inversely correlated (incre ase), and the p H, which did not vary.

Plant Behavior. The tra nsfer of soil-grown Chrysanthe- m u m cinerariaefoliu m plants in hydroponic culture was obtained without plant loss. The growth of plants developed on wastewater and on Co¨ıc-Lessain t solution were identical (Table 3).

The chloroph yll a fluoresce nce was used as a physiological refere nce parameter to follow a possible str ess. Figure 6 shows that the chlorophyll a flu oresce nce of the seedlings of Chrysanthem u m cinerariaefoliu m cultivated on wastewater are not significantly differe nt from those of seedlings cultivated  on  nu tritive solution.

The interest in Chrysanthe m u m cinerariaefoli u m cultiva- tion is its ability of produ cing the insecticidal substances, called pyret hrins, in the capitula. In mosquito coils, pyret hrins I are known to  have  a  goo d  knock-down  effect  while pyret hrins II have a better killing effect (39). Bhat reporte d that the insecticide qua lity of a Chrysanthem u m cinerari- aefoliu m extract is related to two parameter s:  the  tota l conte nt of pyret hrins and the pyret hrins I / pyret hrins II ratio (40). The insecticidal activity of the plant extract incre ases with the pyret hrins I / pyret hrins II ratio. The produ ction of pyret hrins are given in Table 3. The analysis of the tota l

observed that the plant capitula had the same pyret hrin conte nts and the same pyret hrins I / pyret hrins II ratio (data not shown).

The system of wastewater pu rification with NFT using Chrysanthem u m cinerariaefoliu m made it possible, as early as the first month of installation and throughout the following

6 months, to achieve the per mitte d levels for discharge, defined by the Europea n direct ive 21 / 05 / 1991, after 24 h of processing. During this period, the root systems were not saturated , and there was no sedimentation in the chann el botto m. The wastewater see med to provide the necessary elements for growth and a  normal  physiological activity (ph oto synthetic activity  and synthesis  of  seco ndary me- ta bolites). In this study, we demonstr ate  that it is possible to pu rify wastewater, while at the same time produ cing pyret hrins, without affecting either the qua ntity or the qua lity of the capitula extracts .