Table 3. Probit analysis of topical exposure of Temprid SC-selected and unselected lines to the active ingredients (A.I.), β-cyfluthrin and imidacloprid.
| A.I. | Strain | Treatment | n | Slope (±s.e.m.)a | LD50 ng (95% CI) | LD50 ratio (95% CI)b | χ2 (df) | |
|---|---|---|---|---|---|---|---|---|
| β-cyfluthrin | Selected/Unselected | Goodness-of-fitc | Parallelismd | |||||
| CIN1 | Unselected | 480 | 0.27 ± 0.15 | 0.8 (0–403,148) | 51,458 (1,304–20.3 × 105)* | 2.90(2) | 3.90(1)* | |
| Selected | 480 | 0.53 ± 0.10 | 41,518 (9,043–1,147,076) | 1.24(2) | ||||
| LA1 | Unselected | 478 | 1.39 ± 0.13 | 1.3 (1.0–1.7) | 23.8 (15.1–37.4)* | 0.56(1) | 4.57(1)* | |
| Selected | 479 | 1.05 ± 0.20 | 30.5 (4.3–1,437) | 2.74(2) | ||||
| NY1 | Unselected | 660 | 0.55 ± 0.16 | 315.0 (17.8–5,562) | 15.9 (7.3–34.5)* | 7.34(2)* | 9.34(1)* | |
| Selected | 480 | 0.96 ± 0.12 | 5,005 (3,104–9,956) | 2.27(1) | ||||
| Imidacloprid | ||||||||
| CIN1 | Unselected | 480 | 1.04 ± 0.11 | 103.6 (70.1–169.2) | 0.7 (0.4–1.3) | 0.35(2) | 0.31(1) | |
| Selected | 480 | 0.96 ± 0.10 | 71.6 (48.3–115.8) | 1.73(2) | ||||
| LA1 | Unselected | 480 | 1.24 ± 0.11 | 8.0 (6.0–10.8) | 1.8 (1.1–2.8)* | 1.95(1) | 0.67(1) | |
| Selected | 480 | 1.11 ± 0.11 | 14.1 (10.2–20.4) | 0.01(1) | ||||
| NY1 | Unselected | 460 | 0.93 ± 0.12 | 45.3 (23.0–73.1) | 2.8 (1.5–5.3)* | 0.06 (1) | 0.43(1) | |
| Selected | 660 | 1.02 ± 0.19 | 126.2 (8.2–1,522) | 4.81(2) | ||||
*P ≤ 0.05
aSlope, dose that kills 50% of the population (LD50), and goodness-of-fit were calculated using BioStat [AnalystSoft Inc. BioStat v2009 - Statistical analysis program. (2009)]. LD50 ratio and parallelism tests were performed using PoloPlus [Robertson, J.A., Preisler, H.K. & Russell, R.M. Polo Plus: Probit and Logit analysis. (LeOra Software, Petaluma, CA, 2003)].
bLD50 values are significantly different between unselected and selected lines if 1 does not fall within the 95% confidence interval for the ratio test [Robertson, J.A., Preisler, H.K. & Russell, R.M. Polo Plus: Probit and Logit analysis. (LeOra Software, Petaluma, CA, 2003)].
cLarger values of χ2 for goodness-of-fit indicate a poorer fit on the probit regression line.
dParallelism challenges the hypothesis that the slopes are identical for selected and unselected lines [Robertson, J.A., Preisler, H.K. & Russell, R.M. Polo Plus: Probit and Logit analysis. (LeOra Software, Petaluma, CA, 2003)].