332 MODIFIED OIKCUMNUTATION. Chap. VII 



to avoid errors of parallax, all readings were made by looking 

 through a small ring painted on the vertical glass, in a line 

 with the joint of the leaflet and the centre of the graduated arc. 

 In the following diagrams the ordinates represent the angles 

 which the leaflet made with the vertical at successive instants.* 

 It follows that a fall in the curve represents an actual dropping 

 of the leaf, and that the zero line represents a vertically de- 

 pendent position. Fig. 133 represents the nature of the move- 

 ments which occur in the evening, as soon as the leaflets begin 

 to assume their nocturnal position. At 4.55 p m. the leaflet 

 formed an angle of 85° with the vertical, or was only 5° below 

 the horizontal; but in order that the diagram might get into 

 our page, the leaflet is represented falling from 75° instead 

 of 85°. Shortly after 6 p.m. it hung vertically down, and had 

 attained its nocturnal position. Between 6.10 and 6.35 p.m. it 

 performed a number of minute oscillations of about 2° each, 

 occupying periods of 4 or 5 m. The complete state of rest of 

 the leaflet which ultimately followed is not shown in the dia- 

 gram. It is manifest that each oscillation consists of a gradual 

 rise, followed by a sudden fall. Each time the leaflet fell, it 

 approached nearer to the nocturnal position than it did on the 

 previous fall. The amplitude of the oscillations diminished, 

 while the periods of oscillation became shorter. 



In bright sunshine the leaflets assume a highly inclined de- 

 pendent position. A leaflet in diffused light was observed rising 

 for 25 m. A blind was then pulled up so that the plant was 

 " brightly illuminated (BE in Fig. 134), and within a minute it 

 began to fall, and ultimately fell 47°, as shown in the diagram. 

 This descent was performed by six descending steps, precisely 

 similar to those by which the nocturnal fall is effected. The 

 plant was then again shaded (SH), and a long slow rise occurred 

 until another series of falls commenced at BR', when the sun 

 was again admitted. In this experiment cool air was allowed 

 to enter by the windows being opened at the same time that 

 the blinds were pulled up, so that in spite of the sun shining 

 on the plant the temperature was not raised. 



The efiect of an increase of temperature in diffused light is 



* In all the diagninis 1 mm. in ment. In Figs. 133 and 134 the 



the horizontal direction represents temperature is represented (along 



one minute of time. Each mm. the ordinates) in the scale of 1 



in the vertical direction repre- mm. to each 0-1° 0. In Fig. 



Bents one degree of angular move- 135 each mm. equals 0-2° F. 



