It is a common lament amongst agricultural economists that there is a lack of suitable data on input-output relationships in farming, for use in comparative analysis, budgeting, programme planning, linear programming and other planning techniques. However, this does not deter some from inventing the data so that they can use the more sophisticated techniques to reach much-publicized but probably spurious conclusions. Sometimes resort is made to data obtained from experimental farms and there is a lot to be said for this approach. The management factor, which so often bedevils analysis of data obtained from farm surveys, is presumably constant and certain other factors which vary from farm to farm can be kept constant or their variation accounted for in proper experimental design. Thus results can be obtained for input-output relationships (though often at only one or two discrete levels) under stated conditions and with a given farming technique. There is, unfortunately, the problem of translating the experimental results into those that could