Sir, - Not for the first time lately, your Science Today columnist, Dr William Reville, has misled your readers concerning the essential features of the scientific method. In his otherwise excellent article of March 17th, he again maintains that the scientific method necessarily proceeds by hypothesis followed by experiment.

This is a common fallacy or over simplification amongst physical and biological scientists, many of whom seem oblivious of the fact that many major scientific breakthroughs have not involved experiments. I've met biologists who insist that it's the experimental method which differentiates a science. They overlook the fact that perhaps the greatest advance in their own field, Darwinism, was survey based and not experimentally derived.

A further inconsistency in their position is that few of them would deny that astronomy is perhaps the most exact of sciences. Its success in predicting events is based, however, not on experiments but on other methods of observation.

An experiment consists essentially of introducing a change into a carefully controlled subset of the real world, and observing the consequences. Today, many experiments are carried out on mathematical models of the real world, using computers. This procedure is known as "simulation" as widely practised in social and economic studies as in the natural sciences. In other cases, the change may have been introduced not by the researcher, but by some historical event. Yet observation of the outcome may, in certain circumstances, equally enable testing of an hypothesis.

For various reasons - barriers to the use of human beings as experimental objects or difficulties in keeping with other variables constant - some fields of enquiry make much less use of the experimental method than others. Arbitrarily favouring one method of gathering observations over others, and representing it as the unique badge of scientific endeavour, sometimes reflects mere pretentiousness on the part of those in the natural sciences. One would not attribute such snobbery to Dr Reville.

Too often, however, efforts to withhold recognition as sciences from those fields in which experiments are seldom practicable mask an attempt to affect the allocation of research resources between disciplines or the salary differentials between their workers. As such, it becomes a particularly baneful pretension, which should not pass unchallenged. - Yours, etc.

Emeritus Professor of Agricultural Economics,

Coliemore Road,


Co Dublin.