I20 Floivcrs and ihcir Pedigrees. 



extent usurped all the best and most profitable situa- 

 tions in nature. Amon^ them were the immediate 

 ancestors of the goose grass, which had then regular 

 long tubular blossoms, instead of having a mere flat, 

 disk-shaped corolla like the one you sec in the goose- 

 grass before you. But, for a reason which I will pre- 

 sently tell you, in the goose-grass tribe itself the tube 

 hiis gradually become shorter and shorter again, till 

 at last there is nothing left of it at all, and the corolla 

 consists simply of four spreading lobes slightly joined 

 together by a little rim or margin at the base. 



How do we know, you ask, that the goose-grass is 

 descended from such ancestral flowers having a long 

 hollow tube ? Why may it not be an early form of 

 tubular blossom, a plant which is just acquiring such 

 a type of flower, rather than one which has once pos- 

 sessed it and . afterwards lost it.' Well, my dear sir, 

 your objection is natural ; but we know it for this 

 reason. I told you some time since that the other 

 great branch of the madder family, which had stipules 

 instead of whorled leaves, was thereby shown to be a 

 more primitive form of the common type than the 

 stellate tribe, in which these stipules have developed 

 into full grown leaves. Now, all these tropical 

 madderlike plants have large tubular blossoms, per- 

 fectly developed ; so that we may reasonably infer 



