78 Newcombe, Thigmotropism of terrestrial roots. 



and over con.siderable oxtent of root-surface must be applied. 

 Certainly no device can be contrived to meet tlieso conditions 

 so conipletely as a streain of water, tlie pressure of whieh is 

 feit over a füll half of the hypotlietically sensitive area, and 

 instantly accomodates itself to all irregularities of root-surface. 

 But tlie stream of water, if used, must not meet tlie roots di- 

 rectly; otlierwise the responses which might follow would be 

 rheotropic, and we are not yet ready to define rheotropism as 

 thigmotropism, though this lias been suggested. 



The water-stream may however be used, to give pressure 

 in the study of thigmotropism by enclosing the roots in coUo- 

 dion stoclvings. These stockings are made as foUows: To cane 

 sugar are added a little glucose and water, and the mass is 

 boiled tili it reaches a temperature of 110 ^'C After cooling 

 somewhat it is pulled for a while tili it begins to stiften, when 

 small pieces are roUed between the hands into small cylinders, 

 which are finally shaped by rolling between glass-plates , the 

 iinished form being 4 to 5 cm long and 3 mm in diameter. 

 When these cylinders have thoroughly hardened , they are dip- 

 ped into a very thin collodion Solution, one end not being immers- 

 ed, held in the fingers for a minute to aUow the collodion 

 to dry, and then laid aside in a dry place to give the collodion 

 membrane opportunity to become as thin as possible. Next these 

 objects are put into a jar of running wator, and are wholly 

 freed from the sugar in 15 to '60 minutes. 



Before relying on these collodion stockings to shield the 

 roots completely from the flowing water in which the covered 

 roots are to be immersed , the stockings must be tested to see 

 whether they are free from holes, and to determine whether the 

 flow of the suiTounding water could induce a ciu'rent through 

 the collodion of sufficient strength\) to ascribe the curves to 

 rheotropism, should any curves occur. For this fest, glass-tubes 

 of small diameter were heated and drawn out, making them 

 small enough at one end to insert into the collodion stockings. 

 The glass-tubes having been inserted, the edge of the collodion 

 was sealed to the glass-tube by hot paraffin. The collodion cyl- 

 inder having been sunken below water, and the glass-tube pro- 

 jecting above the water, it was found that the collodion tubes 

 were caj)able of sustaining a column of water in the glass-tubes 

 considerably above the level of the surrounding water. Not un- 

 til the water in the glass-tubes was several centimeters above 

 that of the surrounding water, could the downward movement 

 of the water in the glass-tubes be seen. From this behavior the 

 collodion tubes were assumed to be without perforations. 



To fest the filtration power of the collodion walls, H repre- 

 sentative tubes were selected, into the open end of each of which 



1) Jiiel, in Jalirb. f. wiss. Botan. XXXIV. 507, suggested the possi- 

 Lilitjf of a flow of water througli an enveloj)ing coUochou tuhe. 



