activity enabled the plant problem to be solved rapidly .
But any attempt to generalize the experimental observa- tions based on this hypothesis would be filled with un- certainty . If, by contrast , a longer range program of investigation would have been possible , a more creative type hypothesis could have been proposed . Since the structure of the solid catalyst and changes occurring on it during reaction are poorly understood , concepts by postulation would have to be proposed as a basis for any such hypotheses . Con- cepts such as the nature of the support , the mode of ad- sorption of reactant on the discrete nickel crystals , and the geometry of the pores would suggest relationships be- tween variables which could be stated as hypotheses . These hypotheses , when confirmed , would suggest other experiments and finally permit valuable generalizations which are productive of new ideas and inventions and possess great predictive power .
Oncea relationship between variables is established
with some acceptable degree of probability , the powerful tool of mathematics may be used to predict , design , and create with the minimum of additional laboratory work . If, however , just a causal relation has been discovered , only that relation is definitely known , and usually only the particular problem in question can be solved .
The ambiguity in "usually " is intentional since the re-
markable but rare so - called " accidental " discoveries will continue in all fields . These seem to happen to the keenest experimentalist whose power of observation is far superior to the average . He detects a rubbery mass in a reaction vessel and proceeds to develop causal hypotheses and con- firms one by determining the cause . In so doing he invents a new polymerization process . One cannot help but won- der if these astute workers do not also employ more imag- inative hypotheses than simple cause and effect . Imag-