178 bacteria in relation to plant diseases. 



Value of Cultural Characters. 



Of what worth are the cuhural characters commonly mentioned in descripti\'e 

 bacteriology ? Much depends on the proper answer to this question. There are 

 undoubtedly two extreme views, neither of which is correct. One investigator 

 would maintain that no dependence can be placed on them ; another seems to have 

 no suspicion of any source of uncertainty. The truth undoubtedly lies somewhere 

 between the two. That great progress in bacteriology has come from their use 

 must be admitted by all. To cast doubt on everything already done is only to 

 bring chaos back again. It is wise to make haste slowly. No necessity exists for 

 making a rubbish heap of the past before beginning one's own work. Old methods 

 should be tried repeatedl)', scrutinized from every standpoint, and only abandoned 

 when they lia\'e )'ielded all that can be obtained from them, or when there is some- 

 thing distincth" better to take their place. New methods should be hailed with 

 enthusiasm onh' in so far as they have actually made good their claim to be genuine 

 improvements. A great deal of writing on bacteriology is worthless because not 

 based on well-considered and properly conducted experiments. Hypotheses ad 

 libitum, the more the better ; but let us not forget to test each one in the crucible 

 of experiment, and generalh' before publishing^ rather than after. In other words, 

 give to the world only the well-established facts. As a means toward arriving at the 

 truth, let each person not only experiment as carefully as possible, but let him set 

 down all the steps in his procedure, so that others may repeat his experiments. Many 

 misapprehensions and supposed contradictions arise from the fact that workers are 

 led to believe they have exactly duplicated another man's work when they have done 

 nothing of the kind. The temperature at which they have worked has been dif- 

 ferent, or some other physical or chemical condition, important but not recognized 

 or not recorded b)- the first writer, has been unlike, and the results are not the 

 same. Bacteria are not so simple as they appear. While monotonous morphol- 

 ogically they are complex in their multitudinous physiological activities, and are 

 extremely apt to var)' under a slightly changed environment. When we repeat an 

 experiment we must know, therefore, whether we have preserved substantially the 

 former environment. If we have not, then it should not surprise us if the results 

 are somewhat different from those we anticipated. 



A very frequent source of error in interpreting descriptions consists in not 

 making sufficient allowance for changes due to slight variations in the culture-media. 

 I can perhaps make my meaning jjlainer in the following way : Let the curve A B 

 represent all the variations in color and appearance of a given organism on a given 

 medium, e. g., steamed ]3otato. Now, if a worker describes his organism from a 



