32 A CONSIDERATION OF THE EVIDENCE [CH. in 



to revert more or less tjuickly to the original type on the 

 removal of the modifying influence indicates racial stability 

 in character and minimises the significance of the modification. 

 This aspect of the problem will be referred to later {vide p. 144). 

 It is necessary, however, to emphasise at this point the danger 

 of assuming a change in character to be permanent because 

 reversion has not occurred within a certain period, even a 

 lengthy one. A strain of B. ruber, for example, may show no 

 trace of colour for 12 months together under certain conditions 

 and yet retain undiminished its power to produce pigment 

 in more favourable circumstances (Laurent, 1890). In other 

 cases reversion occurs without any modification in the con- 

 ditions of growth but apparently spontaneously and this after 

 long periods of time have elapsed. 



6, The necessity for perseveroMce in following a particular 

 line of investigation needs no less emphasis. Twort (1907) 

 took two years to train a particular strain of B. typhosus to 

 ferment lactose — a result which Penfold (1910 a) failed to 

 achieve in the case of over a dozen strains after a 15 months' 

 trial. Coplans (1909) grew a strain of B. tetcmi on a gelatin 

 medium for 90 days before liquefaction occurred. Eyre and 

 Washbourn (1899) found that to raise a particular strain of 

 avirulent saprophytic pneumococci to full virulence by animal 

 "passage" no less than 53 successive inoculations were 

 required. Goodman (1908) in his attempts to modify by 

 artificial selection the acid production in a strain of diphtheria 

 bacilli, made 18 transfers before any result was perceptible. 



In all these cases, if the experiments had concluded earlier, 

 negative results might have been obtained and a claim based 

 on this evidence for stability in character which a more 

 prolonged investigation would have shown not to be justified. 



7. Faultless technique is essential to accuracy in results. 

 In carrying out agglutination tests, for example, the utmost 

 care and patience is demanded even from a practised observer 

 if his conclusions are to be of any real value, and much of 

 the confusion which at present exists on the subject of 

 agglutination is no doubt to be attributed to bad workman- 

 ship. 



