34 
THE AMERICAN BOTANIST. 
ally herbaceous, the stems are generally hollow, fluted or stri- 
ated, so as to resemble Doric columns. Sometimes the stems 
are knotty, fistular, or pithy. The leaves, always alternate, 
are dilated below into sheathing petioles, familiar to us in cel- 
ery, where it is these, that by bleaching are made palatable and 
even delicious. The blade is rarely entire, usually more or 
less compound or dissected either pinnately or palmately. 
It is from the inflorescence that the family derive its name. 
The flowers are borne either in simple or compound umbels, 
with or without involucres and involucels, or else, much more 
rarely in heads, as in Eryngium, the peculiar button snake- 
root of our western plains. These, with their yucca-like as- 
pect, scarcely suggest the family and are sure to be a surprise 
to the eastern collector seeing them for the first time. 
The individual flowers of the family have the calyx ad- 
herent to the ovary, and either with five very minute teeth or 
none. Then comes an epigynous disk, outside of which are 
inserted the five petals, acuminate, but with the tips so inflexed 
as to give them an obcordate appearance. The range of color 
is limited to white, yellow, blue, or lavender, as a rule, and all 
the flowers are small. In some cases, to a degree in our com- 
mon carrot, the peripheral florets are somewhat larger than 
those of the disk and may be abortive. Carrot most com- 
monly has one central flower of a deep maroon color, giving 
the effect of an insect alight upon the umbel. The function 
of this solitary floret is not definitely known. It may be an 
enticement or guide for small flies. 
It is from the fact that the flowers are so nearly alike 
throughout the entire group, that they fail to furnish definite 
classification features, hence, the fruit and seed only are to be 
certainly relied upon in identification. These furnish an inter- 
esting puzzle to the student. The writer has never considered 
it fair to dismiss them ''as too diflicult for the beginner." Pos- 
sibly that is true for children's classes, but long experience has 
