CLASSIFICATION AND FUNCTION 123 



I think Henslow gave me a Primrose in which the ovules 

 were basal (as normally they should be) in most flowers, 

 and they were parietal in others. It was otherwise 

 monstrous. 



I was much struck with your conclusion that the near 

 approach to uniformity in an organ throughout a group 

 implied its functional inutility it is no doubt true. I had 

 a sort of gleam of this truth when considering the fact you 

 once pointed out to me, that the calli of Oncidium, though 

 essential to the plant for physiological purpose, are still so 

 very variable. It then suggested the converse which you 

 have so well evolved. But what an apparent contradiction 

 it involves or paradox at least that classification and 

 system is founded on the least useful modifications, and 

 this explains a very common observation, that Physiology, 

 i.e. the operations of active plant life, does not much help 

 the systematist. And yet there is something uncomfortable 

 in the idea that system is based on modifications the active 

 exigency of which is no longer in play. It seems frightfully 

 paradoxical to say that the quinary arrangement of Dicoty- 

 ledons is a matter of no moment to the Dicotyledon as such : 

 and yet that this is true is proved by the fact that such 

 Dicots. as are ternary or quaternary are as good Dicots. 

 as their quinary brethren. It is a tremendous upset to 

 Owen's doctrines, or rather his writings, for these in no 

 way rise to the dignity of doctrines. The ' law of necessary 

 correlation ' is nowheres. 



Monday (January 1869). 



Just one last thought anent Genetic characters of no 

 value to the plant : is not the fact, that characters of primary 

 value in system are so often of no use, an argument in favour 

 of your conclusion, that such characters as are of no use, if 

 not in any way detrimental, are not necessarily eliminated 

 but may be retained ad infinitum ? 



On the other hand, is it not an argument against the 

 theory of characters acquired by the individual being heredi- 

 tary thus, if hereditary modifications that never come 

 into play do not die out, is it likely that non-hereditary 

 modifications brought into play by the individual (for its 

 own special use) should be transmitted ? 



