12 



BULLETIN OF THE LABORATORIES OF 



floral axis. The perfect symmetry and alternation of the flower 

 is also maintained in this extra circle. It will be observed that 

 the included stamens are placed alternate to the lines of dehis 

 cence or the midrib. 



The three interior stamens were in 

 every way as perfect as those of the nor- 

 mal circles and their time of dehiscence 

 was the same as the others. The ap- 

 pearance of this flower is shown in Fig- 

 ure 9, without the perianth. Only two 

 of the outer whorles of six stamens are 

 shown, but the three arising from the Fig. 8. 



center of the ruptured pistil and op- 

 posite each valve will give some idea 

 of the peculiar appearance presented 

 by the flower and expressed in the 

 diagram. 



Why the prolongation of the torus 



should have found expression in the 



production of androecium instead of 



a second whorl of carpels will prob- 



^^' ^' ably be explained when the mystery 



which produced the abnormality itself is solved. This much 



however seems quite certain that the ovules of the tulip and 



probably of Liliaceae are carpellary and marginal. 



