Part of this rationale was based on acommitment to a conservative, consistent analyti-cal approach. Another related reason stemmedfrom the high variability exhibited in our systems.Average coefficients of variation calculated forthe control and 200 #g I - dose level chironomiddata were above the range considered acceptablefor replication (10 to 30%, Giddings &Eddlemon, 1979).The analysis of variance (ANOVA) repeatedmeasures was selected over the analysis ofcovariance (ANACOVA) repeated measure as atool for comparing biotic component characteris-tics between the 0- and 200/Mgl-1 treatments.This position was taken since an important con-dition of the ANACOVA was not met, namelythat the slope of the post-treatment data on thecovariate (i.e. pre-treatment primary production,chironomid density, or chironomid biomass) issignificantly different from zero. No biotic com-ponent met this condition (lowest probability wasp < 0.276). Thus, the most powerful test is theANOVA repeated measures (on adjusted data).Phytoplankton primary production was nega-tively correlated with tebuthiuron concentrationat 42-, 57-, and 64-DAT (Table 3). These datesessentially corresponded to peak primary produc-tion as exhibited in the control mesocosms duringthe experimental period (Table 4). There were nocorrelations at any of the other nine post-treat-ment dates. Conversely, when comparing primaryproduction effects of the replicated control and200 Mg I ' treatments, no differences occurredover all time (Table 5) or at any date (Table 6).This discrepancy may be a result of an effect-threshold level located somewhere between the200 g 1- 1 and 500 ig 1 ' nominal doses. Thisconclusion is supported by the Price et al. (1989)tebuthiuron study, wherein 180 Mg 1- tebuthiu-ron did not depress algal production in 400 mlmicrocosms. Moreover, the only detectable effect Table 5. Primary production, chironomid density and biomass, and fish density in mesocosms receiving six nominal doses oftebuthiuron. Means were generated from adjusted post-treatment values (post-treatment value minus pre-treatment value) andare representative of the entire experimental period. Probability values are listed for replicated 0 pg 1 - ' (control) and 200 pg 1 -treatment comparison.Nominal n Primary production Chironomid density Chironomid biomass Fish biomasstebuthiuron (X O2 mg- ' hr-') (X # m- 2) (X g 2) (X g mesocosm ')dose (g - ) (p-value) (p-value) (p-value) (p-value)0 3 0.27 9433 0.9376 65.440 3 0.15 (0.499) 6015 (0.027) 05585 (0.293) 6296 (0.706) 200 3 0.15 6015 0.5585 62.9610 1 0.19 9288 0.8587 72.8570 1 0.21 9161 1.6149 65.55500 1 0.02 4688 0.2880 72.801000 1 0.09 6052 0.5001 53.79in that study was a slight shift in algal communitystructure. Toxicant-produced shifts in assem-blage structure without concomitant depressionin production may be a common phenomenon(Ryther, 1970; Ferens & Beyers, 1972; Couch &Nelson, 1982; Moly & Ruber, 1983).Chironomid density and biomass were nega-tively correlated with tebuthiuron concentrationat 64- and 106 DAT (Table 3). As observed withprimary production, negative correlations devel-oped during periods of elevated standing crop(Table 7). In addition, chironomid density wassignificantly lower in the 200 ptg 1-' treatmentcompared to the control (Table 5). This differenceprimarily was a consequence of substantiallylower density at 64 DAT. This finding is similarto that noted in other toxicological studies (Bowl-ing et al., 1980). Only during peak macroinverte-brate density of the control systems were differ-ences between treatment densities apparent(Bowling et al., 1980). Since density differencesbetween treatments can be time dependent, theresearcher should observe the experimental sys-tems over an extended period. This will helpassure the detection of treatment effects if theyoccur.Although chironomid biomass did not differbetween the control and 200 Ig 1-' treatmentsTable 6. Primary production, chironomid density, and chironomid biomass in the control and 200 pg 1- nominal tebuthiurondose treatments. Probability values generated from ANOVA repeated measures analysis. Means were derived from adjustedpost-treatment values (post-treatment minus pre-treatment value). DAT = days after treatment. (Table 6), a clear trend of lower biomass in the200 #,g 1- treatment was evident. We suspectthat there are biomass effects at the 200 jig 1- level, as supported by a correlation betweenchironomid biomass and density in all meso-cosms (r = 0.65, p < 0.043).We surmise that depressions in chironomiddensity with increasing tebuthiuron concentrationwere due, in part, to reduced primary productionand/or an algal species shift. The strong correla-tion of chironomid density with primary produc-tion (r = 0.74, p < 0.010) supports the first sup-position. In addition, Canton & Chadwick (1983)found that multi-plate artificial substrate samp-lers, similar to those used in the present study,selected against the insect scraper guild (periphy-ton and vascular-hydrophyte herbivores) and fa-vored filterers and gatherers (having diets thatmay contain proportionally greater amounts ofphytoplankton). Tebuthiuron对水产品生产力的影响英文文献和中文翻译(3):http://www.751com.cn/fanyi/lunwen_39733.html