CH. Vl] AUXANOMETERS. 137 



the tissues the stem may be protected by a strip of 

 gummed paper wrapped rouDd it before the cord is 

 attached. Any shrinking or swelling of the earth in the 

 pot will obviously introduce errors. These can only be 

 avoided by watering the plant thoroughly at the beginning 

 of the experiment, and leaving it unwatered during the 

 rest of the observation ; the shrinking of the earth will 

 thus be spread over a considerable period. The most 

 serious error however is the curvature of the stem, which 

 may be either spontaneous or more frequently due to 

 heliotropism. If the plant can be illuminated from 

 above, so much the better ; if not a large bright mirror 

 must be placed close behind it which neutralises one-sided 

 illumination. In our experiments we use principally the 

 flower scape of Narcissus, which is but slightly affected by 

 lateral light. 



(177) The descent of the weight measured on a scale. 



The simplest plan is to fix an index to the weight and 

 read its movement on a vertical mm. scale. A piece of 

 sheet lead 15 mm. x 20 mm. folded across the middle will 

 serve as a weight, and a fine sewing needle placed horizon- 

 tally in the fold of the lead can be secured in its place by 

 hammering the lead. In this way the growth can be 

 estimated in 0"1 mm., and the arrangement might be used 

 for measuring the daily and nightly growth of a plant for 

 a series of days. 



(178) Micrometer screw. 



The weight in this case bears a vertical instead 

 of a horizontal needle, and its descent brings the needle 



