278 MODIFIED CIKCUMNUTATION. CHAT. V. 



upwards, so as to occupy the same position relatively to the 

 upper part ot the main peduncle as in Tr. repens. This fact 

 alone would render it probable that the movements of the sub- 

 peduncles in Tr. repens were independent of geotropism. Never 

 theless, to make sure, some flower-heads -were tied to little stirka 

 upside down aud others in a horizontal position ; their sub- 

 peduncles, however, all quickly curved upwards through tlio 

 action of heliotropism. We therefore protected some flower- 

 heads, similarly secured to sticks, from the light, and although 

 some of them rotted, many of their sub-peduncles turned very 

 slowly from their reversed or from their horizontal positions, 

 so as to stand in the normal manner parallel to the upper part 

 of the main peduncle. These facts show that the movement is 

 independent of geotropism or apheliotropism ; it must there- 

 be attributed to epinasty, which however is enecked, at least as 

 long as the flowers are young, by heliotropism. Most of the 

 above flowers were never fertilised owing to the exclusion of 

 bees ; they consequently withered very slowly, and the movements 

 of the sub-peduncles were in like manner much retarded. 



To ascertain the nature of the movement of the sub-peduncle, 

 whilst bending downwards, a filament was fixed across the 

 summit of the calyx of a not fully expanded and almost upright 

 flower, nearly in the centre of the head. The main peduncle 

 was secured to a stick close beneath the head. In order to see 

 the marks on the glass filament, a few flowers had to be cut 

 away on the lower side of the head. The flower under obser- 

 vation at first diverged a little from its upright position, so as 

 to occupy the open space caused by the removal of the adjoining 

 flowers. This required two days, after which time a new tracing 

 was begun (Fig. 124). In A we see the complex circumnutating 

 course pursued from 11.30 A.M. Aug. 26th to 7 A.M. on the 

 30th. The pot was then moved a very little to tho right, and 

 the tracing (B) was continued without interruption from 7 A.M. 

 Aug. 30th to after 6 P.M. Sept. 8th. It should be observed that 

 on most of these days, only a single dot was made each morning 

 at the same hour. Whenever the flower was observed carefully, 

 as on Aug. 30th and Sept. 5th and 6th, it was found to be cir- 

 cumnutating over a small space. At last, on Sept. 7th, it 

 began to bend downwards, and continued to do so until after 

 6 P.M. on the 8th, and indeed until the morning of the 9th, when 

 its movements could no longer be traced on the vertical glass. 

 It was carefully observed during the whole of the 8th, and by 



