The several stages in the process of bending are shown in the figure below. The actual bending is due to one sided growth at the portion immediately below the insertion of the two sheathing leaves. Examination of this part of the shoot before and after the curvature by marking it with india ink made this clear and it was further noticed that greater growth was as a rule opposite to the bladed sheath so that this was always lower in the curved inflorescence. A rapid bending is seen when the plant is in its best state of growth and the curvature is at two places, one at the base of the shoot and the other just beneath the bracteal leaves referred to above. The bends also appear as loops owing to the basal portion of the shoot being slow of response. The loops were observed in those cases where there was a delay in the development of the topmost flowers. The mutilation of such flowers brought about a similar result as also the complete removal of flowers just opened or of the inflorescence with unopened flowers a little above the insertion of the bracts. The loops were generally characteristic of plants in a poor state of growth. It is interesting to note that the > shaped curvature is reached on the sixth or seventh day of the opening of flowers.
A close study of the phenomenon as explained above shows that we are really dealing with two kinds of geotropic curvatures (1) the positive geotropism of the inflorescence resulting in its complete reversal and caused by a curvature beneath the insertion of the bracts, and (2) a diageotropism of the lower portion of the flowering shoot which corresponds to the internode of the sympodial vegetative axis and behaves as such
