PREDICTIVE DENSITY THRESHOLDS FOR MANAGING HOUSE MICE IN WHEAT AND SORGHUM CROPS

Outbreaks of house mice (Mus domesticus, L.) cause major crop losses in Australia and Central Asia. The extent of pest damage is theoretically related to the pest density and this leads to the concept of threshold pest density (DT). Prior to the present study the concept of DT had not been proven for vertebrate pests. I parameterized a predictive model of threshold mouse density for triggering control in sorghum and wheat depending on the farm-gate value of the crop and effectiveness of the control method in reducing mouse density. The two models are based on the estimated relationship between initial mouse density introduced to crops at early crop maturation (DI) in rodent-proofed pens and yield loss caused by mice (YL). Yield loss assessment was conducted at harvest and there were 49 days between the introduction of mice and harvest. This relationship was asymptotic exponential: YL increased linearly with DI until possible competition between mice occurred at densities in excess of about 500 mice ha-1 in both sorghum and wheat; and competition limited further increases in YL at densities in excess of 2,133 mice ha-1 in wheat and 3,000 mice ha-1 in sorghum. DT is 89 mice ha-1 in wheat and 178 mice ha-1 in sorghum for triggering aerial zinc phosphide baiting if this method is effective in reducing mouse density by 41.15% (wheat) and 52.58% (sorghum). I evaluated three methods to index/estimate mouse populations in the field and in pens. Population estimators based on capture-recapture data gave biased mouse density estimates and therefore could not be used to monitor DT. A track index was correlated with estimated population size in the field. The index was calibrated with known mouse densities in pens and the calibration provides a systematic relationship to predict mouse density in the field by measuring the track index. The track index was also systematically related back to YL. The demographic data collected from these experiments indicate that the cause of the asymptote in the mouse density/yield loss relationships is self-regulation, possibly through spacing behaviours caused by intra-specific competition for space. The self regulated density was 3,100 mice ha-1, which was independent of the crop type and occurred before mice could exhaust food supply from the crops. I used replicated manipulative experiments in the pens to determine that the numeric response of mice to cereal crops and their capacity to cause yield losses were affected by the sex ratio of the initial mouse populations but not by streptobacillary arthritis, a disease caused by natural bacterial infection. DT and the calibrated monitoring methods developed in this thesis can be used for improved strategies for managing YL in sorghum and wheat crops