Abstract
At Armidale, New South Wales, ant theft rates of pasture seeds at sites that had been topdressed with 10 kg ha-1 seed immediately before a four day observation period averaged 27 per cent less than on areas where no such application was made. Strong regression relationships were established between theft rates of the various seed types and both air and soil temperatures, but no relationship was found between either relative humidity or soil water availability and theft rate. High theft rates recorded in the first 24 hours after seed placement were not sustained over longer periods of time. A relationship between the decline in theft rate and the quantity of seed available was established. A series of predictions have been made of the proportion of seed likely to remain during fine weather following oversowing in the Armidale environment. These predictions incorporate seed type, time since sowing and prevailing air temperatures. It is predicted that during winter 70 to 80 per cent of bare seed and 90 to 97 per cent of coated seed could remain 30 days after oversowing. During summer the comparable figures are 7 to 30 per cent and 22 to 36 per cent respectively.