EXPERIMENTATION 



In biology it is often good policy to start with a modest 

 preliminary experiment. Apart from considerations of economy, 

 it is seldom desirable to undertake at the outset an elaborate 

 experiment designed to give a complete answer on all points. It 

 is often better for the investigation to progress from one point 

 to the next in stages, as the later experiments may require 

 modification according to the results of the earher ones. One 

 type of preliminary experiment is the " pilot " experiment, 

 which is often used when human beings or farm animals are the 

 subjects. This is a small-scale experiment often carried out at 

 the laboratory to get an indication as to whether a full-scale 

 field experiment is warranted. Another type of preliminary 

 experiment is the "sighting" experiment done to guide the 

 planning of the main experiment. Take, for example, the case 

 of an in vivo titration of an infective or toxic agent. In the 

 sighting experiment dilutions are widely spaced (e.g. hundred- 

 fold) and few animals (e.g. two) are used for each dilution. 

 When the results of this are available, dilutions less widely 

 spaced (e.g. fivefold) are chosen just staggering the probable 

 end-point, and larger groups of animals (e.g. five) are used. In 

 this way one can attain an accurate result with the minimum 

 number of animals. 



The so-called " screening " test is also a type of preliminary 

 experiment. This is a simple test carried out on a large number 

 of substances with the idea of finding out which of them warrant 

 further trial, for example, as therapeutic agents. 



Occasionally quite a small experiment, or test, can be arranged 

 so as to get a provisional indication as to whether there is any- 

 thing in an idea which alone is based on evidence too slender 

 to justify a large experiment. A sketchy experiment of this nature 

 sometimes can be so planned that the results will be of some 

 significance if they turn out one way though of no significance 

 if the other way. However, there is a minimum below which 

 it is useless to reduce the " set up " of even a preliminary 

 experiment. If the experiment is worth doing at all it must be 

 planned in such a way that it has at least a good chance of 

 giving a useful result. The young scientist is often tempted 

 through impatience, and perhaps lack of resources, to rush in 

 and perform ill-planned experiments that have httle chance of 



15 



