In the era of nearly total reliance on pesticides, there was general disregard of the fact that even the simplest agroecosystem presents a complicated maze or delicately balanced ecological interactions.

The threat of resistance has been realised, and no doubt will continue to be realised in the future. What then is the cost once resistance has developed?

We believe the success of the modern cotton farmer is becoming more dependent upon his being made aware of, selecting and implementing appropriate research and new technology. This requires a strong link between researcher and grower. It is upon this link that we wish to concentrate our comments within the context of the following questions.

The presentation hereunder is a collation of views of consultants from the Darling Downs and the St George regions. The following points will be considered: (a) Are research needs being addressed by researchers? (b) If so , is research information being passed on to growers by consultants? (c) If not , is the information being passed back to researchers by consultants? (d) Thoughts on how to reduce the research gap - what gaps exist now.

If research gaps exist, we are all at fault. If researchers are thought of as 'being in their own little world', it is a ref lection of poor communication efforts on the part of producers and advisors. I f growers and others are accused of taking little notice of research findings, then researchers need to determine the effectiveness of their communication effort . So, the first message to growers and researchers in the two valleys is to improve the two way flow o f communications .

The need to develop marketing strategies for the Australian cotton crop appears to be imperative. If we are to continue to develop nev.1 cotton varieties. lt is essential to be able to place these into the market profitably. We believe the current marketing experience with Siokra is a critical Illustration of how important a very good marketing strategy is if we are to develop markets that will continue to ensure long term demand for our cotton. The all Important quality criteria must be strictly adhered to. Our suggest ion is tr1at the cotton industry requires a formal marketing organisation to perform several important tasks:

On many occasions, we saw some parts of our farm, produce this 3 bale yield, but not over a whole field. One corner ,or one side, or a top or bottom end, but not all over. There is not good value in growing 3 bales in one corner and 1.5 in the other. There seemed to be no distinct pattern of where these best yields would occur. We tried many different practices including chisel ploughing, ripping, and jumbo-busting, but the variation still remained. By 1981 we had a five year average yield of 1.8 bales/acre.

Throughout this article I will be encouraging less tillage. There are three reasons for this. Firstly, tillage is costly. Secondly, soil structure often is sufficiently good for farming without tillage. Thirdly, soil moisture conditions are seldom right for effective tillage: there is always the danger that tillage will be harmful to soil structure.

Firstly, it is important to define "Minimum Tillage" as it applies to our operation. Our objective in achieving such a system is more correctly described as a "Permanent Bed'' system i.e. we have totally directed our efforts to maintaining the same bed. In discussing this subject and the methods we have tried it is also important to point out that our experience is confined to the heavier soil types of the Lower Gwydir System.

This project was prompted by disappointingly low cotton yields at Emerald and other Central Queensland centres. The aims of the project are to provide a basis for sound irrigation management of cotton through studies of ( i) Crop response to a range of irrigation management options in terms of crop irrigation efficiency (yield per unit water applied) and plant fruiting development. (ii) Soil water balance in relation to irrigation application efficiency (infiltration per unit water applied), crop water use, ground-water accessions and model predictions of all these factors.