An assessment of the impact of biological and chemical grasshopper control agents on grounddwelling arthropods in Niger, based on presence/absence sampling

Medium-scale experimental field applications of two grasshopper and locust control agents, the mycoinsecticide Metarhizium anisopliae (flavoviride) var. acridum (Deuteromycotina: Hyphomycetes) and the organophosphate fenitrothion, were carried out in August 1996 in Maı̈ne Soroa, Niger. The objective was to evaluate effects and side-effects on target grasshoppers and epigeal non-target arthropods, using simple binary (presence/absence) sampling techniques. Plots (≈50 ha, 3 replicates) were treated at standard recommended dose rates of 5 × 1012 spores/ha (Metarhizium) and 250 g a.i./ha (fenitrothion). The frequency of occurrence – or total presence – was used as a measure of relative population density. The fauna was monitored at 5 day intervals for 36 days (5 d before and 31 d after treatment) and again in August 1997, using 40 miniature pitfall traps per plot and sampling date. Presence/absence sampling proved appropriate to detect and quantify pesticide effects, and the use of similar techniques in rapid risk assessment programs (RAPs) for locust and grasshopper control is recommended. Both control agents were effective against grasshoppers, reflecting their inherent mode of action and speed of kill. The efficacy of fenitrothion ranged from 75 to 86% over the 31 d post-treatment period. Metarhizium had an efficacy of 76% during the last sampling period (days 21–31). Differences between treatments were not significant during this phase. The non-target arthropods monitored on species or species group level represented four insect families which together made up about 75% of the total catch: Carabidae, Tenebrionidae, Formicidae and Ephydridae. 75% of these taxa were significantly reduced by fenitrothion. The median effect was 69% (first week), 33% (second week) and 51% (third to fourth week), respectively, and fenitrothion was classified as `medium risk'. Most of the non-target fauna had fully recovered after 31 days, but one ant species was still significantly reduced in 1997. None of the taxa monitored proved susceptible to the mycopesticide. The median effect was