298 PLANT-BREEDING 



surround the capsule by a large crown of slender and green 

 organs, which persist after the petals and stamens are 

 fallen off, and thus heighten the ornamental effect of the 

 fruits. Only the innermost rows of stamens are changed 

 in this way. Both the filament and the anther are affected, 

 the former being dilated into a sheath, in whose cavity the 

 ovules are produced. Ordinarily these latter are only 

 imperfectly developed and the bright crown of secondary 

 pistils contribute Httle if any to the fertiHty of the flower. 



What makes this variety especially suitable for our dis- 

 cussion is the high degree of variabihty in the development 

 of this crown of pistillodous stamens. In some instances 

 it is seen to be so broad as to have affected a large majority 

 of the stamens; in others only one or two of these parts are 

 changed, and in extreme cases only a single hardened and 

 persisting filament without an ovarial cavity and without 

 stigma indicates the type. All intermediate forms may be 

 found and the variabihty covers a range, going from one 

 pistil up to a hundred and fifty and sometimes even more. 

 Unfortunately these altered organs often show another 

 anomaly, consisting in a coalescence of some of them, thereby 

 constituting groups, in which it is difficuk to distinguish and 

 count the single parts. In extreme cases this coalescence 

 may be such as to combine all the changed stamens into a 

 narrow cup around the normal capsule, and more com- 

 monly a more or less divided ring of such groups is observed. 

 It is easily seen that the variabihty of this character affords 

 an excellent material for the study of its dependency on 

 outer influences. We may make sowings under different 

 circumstances, giving them a sunny or a shaded place, good 

 or bad soil, different quahties of manure, watering or keep- 

 ing them dry, protecting the young plants under glass or 

 exposing them to all the effects of the weather, and so on. 

 If we do so we easily find that all favorable conditions 



