COMPLEXITY OF IMPATIENS SPECIES 395 



awake.' If the first systematic difficulties had been overcome, 

 he found that to complete his descriptions he had to analyse 

 most of the species afresh, and even then the descriptions were 

 ' vague and loose,' for ' every organ is variable except such 

 as afford no character at all : it is like making species out of 

 the waves of the sea. Still the species are stable enough, 

 however loose on their pins ' ; and he plodded on, in his pet 

 phrase * groaning over Impatiens ' at the rate of about two 

 species a day. 



New specimens continually arrive, each with its own problem 

 to raise or to solve. One beautifully preserved set from Gamble 

 displays characters that had been destroyed in the more 

 roughly handled herbarium specimens, and * clear up a great 

 long standing puzzle.' ' They settle I. Gardneriana definitely ! 

 but what a queer beast it is ! the total dissimilarity of several 

 states is very striking.' (November 25, 1903.) 



Of one curious Impatiens indeed (I. tingens), he had written 

 to Sir P. Darwin (June 11, 1903) that it ' differs so greatly 

 from all its congeners that it may be called a case of 

 evolution per saltum. 9 



Thanks to another set he makes out in the following June 

 that the I. scabrida of the F.B.I, should be broken up into 

 three species, among which he owed the determination of De 

 Candolle's original scabrida to a photograph of the specimen 

 in the De Candolle herbarium. * I cannot tell you/ he ex- 

 claims, * the worry these three have cost me ; due to the bad 

 specimens, the confusion in Herb. Wallich, and my own care- 

 lessness or stupidity, or both.' 



Long after, Gamble sent him for identification a Balsam 

 which appeared to be I. scabrida. He placed it as one very 

 like this, which had already figured under six names, adding 

 (September 28, 1909) : 



As to your finding that yours does not properly fit the 

 description of either, that is the normal condition of every 

 described Balsam. I am now revising my original clavis 

 of the Chinese species with the help (hindrance I should say) 

 of additional specimens and duplicated analyses, and the 

 operation is most disheartening. 



