l8 9i.] Abnormal Phyllotaxy. 161 



considered, the fact that they can all be reduced to a few 

 series is extremely interesting. It is especially to be noted 

 that the abnormal conditions immediately to be cited, regard- 

 ing the arrangement of the petals, have no subsequent effect 

 upon the position of the pistil leaves, and therefore can not 



• * - -*■ 



consist in the interpolation of new whorls of petals, but must 

 arise from dedoublement and the new conditions which de- 

 doublement in preceding whorls often forces upon those 

 which immediately succeed. The most common case consists 

 in the addition of one, figs. 4 and 5, or two petals, fig. 6, to 

 the inner whorl of petals, in such a manner that they take a 

 diagonal position in the flower diagram, more or less approach- 

 ing the transverse position which a dimerous whorl should 

 take if it were added to the whorls already existing. That 

 these petals are not to be interpreted in this way, but as 

 formed by dedoublement in the nearest petals, is suggested 

 by the unchanged position of the pistil leaves, as already 

 mentioned. A second case consists in the removal of one, 

 fig. 7, or two petals, fig. 8, belonging to the inner whorl of 

 four petals, to the middle dimerous whorl, thus giving the 

 second whorl of petals three or four members instead of two. 

 ■since this removal takes place in a direction away from a 

 transverse line connecting the sepals the result is to leave the 

 original petals in nearly their normal position, and to bring 

 the added petals into a median position. That this second 

 whorl does not consist typically of four petals in these cases 

 is shown by the fact that the space opposite the intercalated 

 petals is left vacant, whereas a whorl of four members should 

 be equally distributed. According as one or two petals have 

 been removed from the inner whorl to the middle whorl, one 

 or two petals are added to those remaining in the inner whorl 

 and these added petals are again placed near a transverse line 

 connecting the sepals, but on the same side of # this line. 



In a third case, one, fig. 9, or both petals of the second 

 dimerous whorl have become two by dedoublement. In that 

 case the four petals of the inner whorl occupy as nearly as 

 may be their original positions and one, fig. 9, or two petals 

 are added to their number and then occupy the interval created 

 by the separation of petals of the second whorl into two 

 through dedoublement. In this case the addition is made 

 precisely in line with the transverse position occupied by the 

 sepals. The three variations from the type just described are 



