STRUCTURE OF FLOWERS. 127 



axis is capable of being prolonged into a branch with 

 leaves, and the same disposition is met with also in 

 flowers where the axis is but little if at all prolonged, 

 provided that the carpels are not arranged in a regular 

 verticil. Thus, flowers with verticillate carpels are true 

 terminations of branches ; those with spirally arranged 

 carpels may be terminations of branches, but are so only 

 by the exhaustion of the central parts, and when these 

 are well nourished, or when the sexual parts are abortive, 

 the branch may be prolonged by its apex. 



The number of parts in each row of a verticil is fixed 

 in each plant, often in each family : it is most frequently 

 quinary in Dicotyledons, and ternary in Monocotyledons. 

 We may believe that the number of parts in the verticils 

 or rows of the same flower is naturally the same in each ; 

 but it appears very different in several cases : — 1st, Be- 

 cause the number of rows of verticils is different : thus, 

 frequently there is one row of petals, and two, three, 

 four, &c. of stamens, &c. 2d, Because there is an abor- 

 tion, union, or metamorphosis of some parts. 



The parts of each verticil or of each row are capable 

 of being united together by cohesion in every possible 

 degree, and the degree of union causes what are called 

 the divisions of the parts. 



The parts of each row of a verticil or of each uni- 

 serial verticil are generally placed alternately with those 

 of the preceding row. When the verticils are multi- 

 serial, each row is likewise alternate with the preceding 

 and following one. The verticils and their rows may be 

 unequal, and not resemble each other, without the flower 

 ceasing to be regular ; it becomes irregular when one of 

 the parts of a verticil or row is different from others of 

 the same row. 



The parts of each verticil are capable of being united 

 by adhesion with those of the neighbouring one ; thus, 



