11 



niicy not extend l)e\oiul its smnmil in an\- form, in whicli case it is 

 said to be obsolete^ or it may appear at the top of the ovary as five 

 persistent teeth or lobes, minute or somewhat conspicuous. The 

 picsiMicc or absence of these *'caly x-lectli"' fuinishes a character 

 used in <^cneric descriptions. 



Petals. The d-jciduous petals are inserted at tiie edj^e of an 

 epigvnous disk, aiul are usually white or yellow, althou<;h in some 

 species they are blue, or even piid<ish to dark purple. L'sually 

 they are concave, with an intlexed ti[) or acumination; and some- 

 times at the flexure tb.e acumination liecomes so impressed as to 

 j^ive an emarginate or 2-lobed appearance. Much use has been 

 made of these characters in generic descriptions, but among our 

 North American forms we have found such uniformity that char- 

 acters drawn from the petals are of no value, witii the single ex- 

 ception of the genus Hcraclcitiii. 



Stamois. From the 5 alternate distinct introrse stamens no 

 characters are to lie drawn. In general, (7//ibc//ifci\c seem to be 

 protandrous, though protgyny has been obsi-rved in a few genera, 

 such as 'riiaspiitni^ Saiiicida and Hrigcii'ia. 



Pistil. A description of the bicarpellary pistil, surmounted 

 bv the disk from which arise two distinct styles eitb.er directly or 

 with an intervening. v/i7^/'f'<//V^w, which is either depressed or more 

 or less conical, is better includeil under the discussion of the very 

 important fruit structures. 



Fruit. — In no family of plants does the fruit furnish more 

 certain diagnostic characters than in the Umbellifene. So definite 

 are they that the fruit alone can be made to determine the genus, 

 and in most cases the species, while in every case it is an essential 

 part of the description. This indicates at once an unusual amount 

 of differentiation ir. the fruit structures, and great diversity in their 

 display. 



The two carpels face each otiicr, antl are in contact at fii'st by 

 their commissural or \entral faces, but eventually separate. The 

 fruit is compressed lateral Iv (at right angles to the plane of the 

 commissural faces), or dorsally (parallel with the commissural 

 plane), or not at all. As the two carpels are but repetitions of 

 each other, a description of structure may be confined to a single 

 one. The surface of the carpel is usually marked bv five ribs or 



