190 HOW TO STUDY PLANTS. [LESSON 31. 



to Class I., ANGIOSPERIMLE, and then to Subclass I., DICOTYLE- 

 DONS. The tubular corollas in our head of flowers at once lead us 

 to Division II., GAMOPLTAL^E (p. xii.). 



550. Under Gamopetalas two choices are presented, marked A 

 and B. In our plant the ovaries are most decidedly inferior, and 

 under A we meet four choices, based on the number of stamens. 

 Examining the tubular flowers, we find at the summit of the tube 

 five small spreading teeth, which indicate that we are dealing with 

 a five-lobed corolla. Carefully slitting the tube we find five sta- 

 mens which cling together by means of their anthers, an arrange- 

 ment called syngenesious (286). All this easily determines a 

 choice of the second section, " Stamens as many as the lobes of 

 the corolla, 5, syngenesious." Under this are two choices, and, as 

 our flowers are "in an involucrate head," we are directed to tho 

 order COMPOSITE, page 129. 



551. It will usually be entirely unnecessary to use this Analyti- 

 cal Key for Composite, as the flowers of this order are so charac- 

 teristic that it can be at once recognized, and the student can turn 

 directly to the family as presented in the Manual, beginning at 

 page 129. 



552. In the " Key to the Tribes " it will be noticed that our ten 

 tribes are grouped in two series, TUBULIFLOR^E and LIGULI- 

 FLOR^E. As our plant has tubular flowers, and the second series 

 demands that all the flowers shall be ligulate, the choice falls on 

 TUBULIFLOR.E, containing nine of the ten tribes. These nine 

 tribes are grouped under two headings, both beginning ' Heads 

 homogamous," etc. In the first, the heads are "discoid" (without 

 rays), and "never yellow," which at once brings us to the second 

 group (p. 130). Under this seven tribes are arranged, character- 

 ized by various combinations of characters. Our anthers are 

 surely not caudate, which character enables us to discard Tribes 

 IV. and IX., in both of which also the heads are discoid. In 

 Tribe V. the receptacle is chaffy ; in Tribe VI. the receptacle is 

 naked, but the pappus is not abundant and soft-hairy as in our 

 plant ; while in Tribe VII. the pappus is far from agreeing with 

 ours. This leaves Tribes III. and VI II., between which the 

 copious capillary pappus, simple involucre, and naked receptacle 

 decide in favor of the latter, making our plant a species of the 

 SENKCIONIDE^E. 



553. Turning to page 139, we find that the supplementary tribe 



