FLOWERS THAT NEVER FADE 17 



were before him, so that he should find the 

 right number of stamens in every flower, 

 the proper degree of pubescence on the 

 stem, and such other characteristics as only 

 a microscopic examination would reveal. 

 He compared each plant with his own 

 herbarium specimens, relying more on the 

 natural plant than on the printed characters. 

 He found that each flower had its separate 

 character, like nature, and that the indica- 

 tions that one flower was older than another 

 were shown, from the young buds through 

 the partly expanded blossoms to the fully 

 opened flowers, as in the living plants. 



" Sixteen species I examined by careful 

 comparison," he tells us, "besides making a 

 more general observation of a large num- 

 ber. I sought faithfully to find some error, 

 something systematically wrong. A fair 

 criticism should disclose whatever faults 

 may exist, but I failed to find such faults. 

 In a specimen cluster from the angelica- 

 tree, with its flowers so small that their 

 structure can be seen only with a lens, while 

 many of its buds are so minute as to be in- 

 distinguishable to the naked eye, I counted 

 of buds, blossoms, and developing fruit, 

 from twenty-five hundred to three thousand. 

 And yet every flower has its five petals, and 

 five alternating stamens with long filaments. 

 I sought to find on the under part of the 



