CREATIVE SCIENCE 337 



Priestley's were. He was then prepared to find some variability in the 

 appearances; and so, understandably, failed to observe the small dif- 

 ferences in the presence of the far more obvious, and presumably 

 significant, similarity. 



Priestley's discovery of the inadequacy of his first identification is 

 also typical: a tacit expectation founded on this hypothesis broke 

 down by a wide margin. Here the discovery is peripheral. Priestley 

 noticed that during storage over water a specimen of the new gas 

 remained substantially unchanged in volume after a period in which 

 laughing gas would have been largely if not completely dissolved. 

 That is, Priestley noticed (though this lay well beyond the focus of 

 his attention) that the new gas is much less soluble than laughing 

 gas. The TI surmounted, Priestley turned back to repeat the combus- 

 tion experiments. And now, having reason to look more closely, he 

 did notice the small diflFerences within the overriding similarity. The 

 new gas is not laughing gas. 



Priestley here emerged only to plunge right back into error. He 

 applied his nitrous air test to the new gas and saw that— like ordinary 

 air— it gives a substantial volume contraction and a copious produc- 

 tion of brown fumes. The similarity is overwhelming, and he had 

 then no doubt about his second ( mis ) identification of the new gas 

 as common air. The proportionate volume contractions are noticeably 

 diflFerent; but the difference does not much exceed the intrinsic varia- 

 bility of the test, and so was not noticed. This kind of error is not 

 peculiar to Priestley: the identical error was made in the identical 

 way by the celebrated Lavoisier. Lavoisier indeed never detected the 

 error until it was pointed out to him by Priestley, who had by then 

 again encountered, here as a focal discrepancy, a major breakdown 

 of an expectation founded on the prior identification. In a random 

 experiment he found that the residue left from the nitrous air test of 

 the new gas still supported combustion brilliantly. But when ordinary 

 air is tested the residue, far from supporting combustion, extin- 

 guishes a flame. The discrepancy is immense, the TI is exceeded, 

 and Priestley returned to a more careful examination of the nitrous 

 air tests. And then at last he noticed the small differences and, ulti- 

 mately, identified his material as a new species of gas. 



This ancient story deserves review only because it is representa- 

 tive of so many other like instances in the history of science. The 

 problem of identification is, of course, still with us— observe the de- 



