Why Flowers Are in Clusters. 97 



umbels, the latter than those in racemes, which, in turn, 

 must surpass those widely separated. 



Let us see if this is not reasonable. Suppose the flowers 

 which form the umbellate cyme of our common garden 

 geranium more properly a pelargonium were placed 

 separately opposite the leaves. At once we see that the 

 pedicels now peduncles would have to be considerably 

 lengthened to bring the flowers distinctly into view above 

 the leaves. This would be an additional expense to the 

 plant. It is plain, also, that the scattered flowers could not 

 be seen as far as the larger signal formed by the umbel. 

 Few^er insects would visit them, and it would take them 

 longer to go from one to another of the separated flowers 

 than to pass from nectary to nectary of those in the cluster. 

 In like manner it can be shown that flowers do better work 

 in a head than in an umbel. In composite flowers of the 

 highest type there is something more than the mere arrange- 

 ment of the florets in a compact head. The florets closely 

 fitted together stand upon their united receptacles sur- 

 rounded by one or more series of protective bracts. The 

 outer row of flowers (rays) usually have large corollas, the 

 petals of which, instead of forming tubes, are united to 

 make flat signal banners more than three times as broad as 

 a tube would be formed of the same petals. The ring of 

 rays contrasting in color with the disk florets makes a 

 target-like signal easily distinguished at a distance. The 

 disk corollas are small nectar cups. The calyx tube, not 

 being needed to protect in the usual way, becomes the chief 



