29 



One whorl of replaced stamens consisted of four pistils of the 

 nature just described and, in place of the fifth, a filament-like 

 object of about the same length as the other pistils and terminated 

 by a stigma. Probably this was a pistil devoid of an ovary, 

 yet its appearance was that of a stamen whose anther had been 

 replaced by a stigma. Another instance very similar to the 

 last consisted of a whorl of four pistils and one stamen in place 

 of the whorl of stamens; and again in an arrangement, likewise 

 alternate with the petals, of eight pistils in four groups of two, 

 and a pistil and stamen-like pistil, as described above, together 

 in the fifth group. 



The petals when replaced by stamens exhibited similar grada- 

 tions. In one flower instead of the normal five petals were 

 found (i) three petals, (2) a filament-like member terminated 

 by a small petal-like appendage, and (3) a filament-like member 

 terminated by a similar petal-like part but with two swellings 

 that suggested the formation of an anther. Another flower 

 exhibited for its whorl of petals three petals, one filament-like 

 petal, and one stamen. Finally, in still another were combined 

 several of these instances. For the whorl of petals were three 

 petals, a filament-Hke petal, and a stamen; for the whorl of 

 stamens were four pistils and one stamen. In no cases were the 

 sepals or the main pistil variable. 



The setting of the fruit was normal and the abnormal pistils 

 were lost when the calyx, from which they arose, was pushed off 

 by the developing main ovary. There were two trees of the 

 variety under observation and both exhibited the same pheno- 

 mena. Whether this peculiarity is a varietal characteristic or 

 the result of a frost which occurred about the time of blossoming 

 is not known. 



N. Y. Agricultural Experiment Station. 



Geneva, New York. 



