THE ART OF SCIENTIFIC INVESTIGATION 



unjustified, and the cause should probably be sought in the 

 poverty and overcrowding which is to be found in these 

 insalubrious areas. Virchow, in refuting Semmelweis' doctrine 

 about the causation of puerperal fever, asserted that the 

 weather played an important part, because the highest incidence 

 occurred in winter. Semmelweis replied that the association 

 between epidemics and winter was due to the fact that it was in 

 winter that the midwifery students spent most time on the dis- 

 section of dead bodies. 



False conclusions can be drawn by attributing a causal role 

 to a newly introduced factor whereas, in fact, the cause lies in 

 the withdrawal of the factor which was replaced. Tests carried 

 out among people accustomed to drinking coffee at night could 

 show that a better night's sleep was obtained when a proprietary 

 drink was taken instead of coffee. It might be claimed that the 

 proprietary drink induced sleep whereas the better sleep might 

 well be entirely due to coffee not having been taken. Similarly, 

 false conclusions in dietetic experiments have sometimes been 

 drawn when a new constituent has replaced another. The 

 supposed effect of the new constituent has later proved to be 

 due to the absence of the article of diet displaced. It was 

 found that the blooming of some plants was influenced by 

 supplementing day light with artificial light. At first this was 

 thought to be due to the prolonged " day ", but subsequently 

 it was found to be due to the shortened " night ", for breaking 

 into the night with a brief period of illumination at midnight, 

 was even more effective than a longer period of illumination 

 near the evening or morning. 



There is always a risk in applying conclusions reached from 

 experimentation in one species, to another species. Many 

 mistakes were made in concluding that man or a domestic 

 animal required this or that vitamin because rats or other 

 experimental animals did, but nowadays the error of this is 

 generally appreciated. More recently the same trouble arose in 

 chemotherapy. The sulphonamides which gave the best results 

 in man were not always found to be the best against the same 

 bacteria in some of the domestic animals. 



A rather more insidious source of fallacy is failure to realise 

 that there may be several alternative causes of one process. 



ii6 



