80 



NORMAL AND EXPERIMENTAL POLLINATION. 



The visitors to normal flowers as observed for an hour on July 5 and 8 

 are given in table 55. There were 50 Monarda heads used on the first date 

 and 80 on the second. 



Table 55. — Visitors and visits to normal flowers. 



As would be expected from the form of the flower, more than half the 

 visitors belonged to the two genera of butterflies, Erynnis and Argynnis. 

 While all four species of Bombus are able to secure nectar in the proper way, 

 this is sufficiently difficult so that they often avoid Monarda when Geranium 

 or Aster are near, or they steal the nectar by puncturing the corolla at the 

 base, as was done by all the visiting individuals of Bombus proximus and 

 morrisoni in the present observations. 



Meehan (1892:449) found that the anthers of Monarda fistulosa burst as 

 the lips expand and while the stigma is still infolded by the curved portion 

 of the upper lip. By nightfall of the second day the stamens begin to wither, 

 the upper anther-cell shriveling first. In the morning the styles grow out 

 beyond the upper lip and the stigmas separate. Keller (1892 : 452) observed 

 that the anthers shed their pollen and the stigmas diverge widely in the 

 closed buds, but this was probably a case of cleistogamy due to cold, as 

 the observations were made in November. Longyear's observations (1909: 

 84) are in accord with those of Meehan, as is the detailed life-history (plate 

 13, figs. 11 to 16). 



EXPERIMENTS. 



MUTILATION AND COMPETITION. 

 Plan. — In order to permit a wider range of selection, mutilation and 

 competition experiments were combined in the case of Monarda. The 

 flowers of this plant lend themselves with especial readiness to studies of 

 the effect of mutilation, owing to the pronounced zygomorphy. Moreover, 

 the arrangement of the flowers in heads affords an opportunity to vary the 

 kind of competition. Paper disks and paper composite flowers were also 

 employed in some of the series. In order to determine in some degree the 

 effect of place and time, two observers were employed to record the behavior 

 in two separate clumps, distant 1 to 3 meters from each other, and the 

 calendars were repeated a week apart and on two successive days. In 

 consequence of the closing of the flowering period for many plants, the 



