Water turnover rate (WTR), urine concentration and field metabolic rate (FMR) were examined in house mice, Mus domesticus, permanently inhabiting roadside verge areas and seasonally invading crops in semi-arid wheatlands in South Australia. FMR was approximately proportional to body mass and mean values varied from 4.8 ml CO gh (2.9 kJ gd) in autumn and winter, to 7.0 ml CO gh (4.2 kJ gd) in maturing crops during spring. WTR was independent of body mass, indicating that larger mice were selecting a diet containing moister foods. WTR was low in summer and high in winter, and in mice from crops varied from 165 ml l body water d (122 ml kgd) to 1000 ml ld (725 ml kgd). Seasonal changes in WTR were less extreme on the roadside, where a greater diversity of food was available. In the crops, breeding occurred throughout summer during two of three years, but the population increased only in the one summer when mice had marginally higher WTR. On the roadside breeding and population growth were continuous during summer, except in a drought year. Avcrage urine concentration was inversely related to WTR, and varied from 2.0 to 4.8 Osm l. The data indicate that the water conserving abilities of mice equal those of many desert rodents. The water conserving abilities of mice living in crops during summer were fully extended, and in some years aridity limited breeding success and population levels. The degree of moisture stress to which mice are exposed during summer appears to depend not only on rainfall but also on other factors such as availability of food and shelter, and the level of weed infestation in crops.

Warren ripping and poisoning were used to control rabbits on the flood-out plain of a major creek system on Manunda Station, a sheep-grazing property near Yunta in semi-arid South Australia. Rabbit numbers were initially reduced by >99 per cent, as indicated by the number of active entrances remaining in rabbit warrens. After nearly 10 years without follow-up control work, ripped warrens had only two per cent of the pre-control number of active entrances. Poisoning effectively reduced rabbit numbers in the short-term, but had no long-term effect on the number of active entrances, either in ripped or unripped warrens. Perennial shrubs regenerated on and around ripped warrens. Warren ripping on this part of Manunda is a cost-effective management option.

Mouse populations were monitored at 15 sites between 1980 and 1990, during which time one severe mouse plague, in 1980, and one minor outbreak, in 1984, were recorded. Smaller annual peaks in autumn to early winter were followed by winter population declines. Crops were colonised each year in late winter or early spring by mice from winter refuge habitats with dense, low vegetation, including roadsides and grassland along a railway line. In most years mouse numbers in crops declined during summer, but in 1983-84 they rose continuously during summer and autumn, and reached very high levels. Crops planted in 1984 were invaded by large numbers of mice which had survived through winter in the paddocks, but population levels again crashed in late spring and summer. Recorded population changes were generally consistent with plague probabilities predicted from environmental variables, except in 1985 when numbers failed to reach the predicted high levels at most sites. Population changes in crops during late spring appear to be critical in the development of mouse plagues. Large litter sizes and pregnancy rates, and variable survival rates and size of the breeding population, appear to be important factors at that time.

Information on the occurrence and distribution of plagues of house mice, Mus domesticus, in South Australia was obtained from scientific literature, agricultural reports, rural and suburban newspapers, and personal observations. In all, 580 separate reports from the cereal-growing areas of South Australia in 20 years between 1900 and 1984 were collated to determine the extent of each plague. Some plagues were widespread, affecting nearly all 22 000 km2 of the cereal-growing areas; others were restricted to specific areas or isolated sites. Maps showing the extent and relative severity of mouse infestations are presented for 20 years between 1904 and 1984.

The grain-growing areas of South Australia were affected by a severe mouse plague during the winter and spring of 1993. Damage to recently sown and maturing crops was minimised by broadcasting 0.3% strychnine-treated wheat across affected crops, at a rate of 1 kg ha-1. Three indices were used to measure relative mouse abundance before and after treatment: counts of active mouse holes, bait card consumption, and live-trapping. Hole counts and live-trapping both underestimated treatment effects. Bait card consumption provided the most accurate indication of treatment effects. In crops treated across their entire area, treatment reduced bait card consumption by 87%, with 95% reduction in 18 of 28 crops monitored. In most cases, baiting stopped damage by mice and allowed farmers to establish healthy crops where previously mice had removed all the seed sown, and resown, prior to treatment. Perimeter baiting was less successful owing to rapid reinvasion of treated areas, and reduced average bait card consumption by only 16%. Strychnine baiting in crop stubbles was ineffective where weed seeds were abundant.