XXVL INTRODUCTION. 
appears to be universally increased upon one plan. In 
some sections the flowers are solitary, as in Pimpinelli- 
folie, and may be known to be so by the absence of 
bracteze, which are not produced till the flowers are 
increased in number. When therefore these are pre- 
sent, the flower, if apparently solitary, will be found to 
have become so by the abortion of other lateral flowers. 
The first approach to composition is by a flower with a 
bractea at its base being produced on each side the 
primordial one. When the inflorescence is again in- 
creased it is by a similar addition of flowers on each 
side of the secondary ones; so that those which were 
lateral with respect to the primordial flower, become 
central as regards those which spring from themselves ; 
and so on. The inflorescence then must always be 
considered to begin where the first central flower, which 
blows the earliest and has the shortest stalk, appears ; 
and therefore all ramifications without it, however 
aphyllous they may be, must be considered as branch- 
lets. Let, for instance, moschata be examined in its 
most compound state. The mass of ramifications is 
found to consist of alternate ramuli, usually furnished 
with a leaf at their axille. Each of these throws forth 
other alternate branchlets; the infra-axillary leaf being 
perhaps reduced to a single pinna. These last are 
again subdivided into fresh branchlets; and if this be 
the ultimate stage of composition, as it usually is, each 
branchlet is terminated by an ebracteate flower; and 
there inflorescence must be considered to commence. 
The primordial flower, as it expands the earliest, is 
probably the most perfect; and should therefore be ex- 
amined to ascertain the number of ovaria, which al- 
though tolerably constant in it, are by no means so in 
