i62 BOTANICAL GAZETTE [September 



about 40^"" long, were filled with plaster of Paris. As the plaster 

 began to set they were put in a dish of boiled water. They 

 were then boiled, cooled, and boiled again, boiled hard altogether 

 two hours, and left in the same water over night. Air could be 

 located by reflection between the plaster and the glass, as isolated 

 bubbles, making on an average probably one-fifth to one-fourth 

 of the area ; the amount of air remaining in this way varied in 

 different tubes. These tubes were connected with heavy rubber 

 tubing; this was done by filling the rubber connections with 

 fresh plaster paste and immediately forcing this out with the 

 pieces of glass tubing to be connected. A continuous tube of 

 plaster of Paris was thus made 12.4°" in height, reaching up the 

 well surrounded by three flights of stairs in the Hull Botanical 

 Laboratory. Four meters below the top was a T tube full of 

 plaster, w^hose free arm was closed. The lower end was con- 

 nected with a U tube full of water and mercury in equilibrium ; 

 this represented the root. At the top, above another T tube full 

 of plaster with the free arm closed, was the leaf — a sort of 

 inverted osmometer made by filling a funnel with plaster contain- 

 ing some K^FeCNg, tieing over it four sheets of parchment 

 paper soaked in K^FeCN^, sealing a glass wall around this, 

 and filling the cup thus formed with a concentrated solution of 

 CuSO^ and NaCl. The surface of liquid exposed was great 

 enough so that in the dry, warm air in the hall evaporation was 

 more rapid than the loss was supplied, so that after the first day 

 there was a slow precipitation, and after several days more water 

 had to be added to forestall drying. 



The experiment was started at 3:18 p. m., January I4- ^ 

 recorded with very numerous measurements the rise of mercury 

 in the closed arm of the U tube, most of which measurements it 

 would be a waste of space to publish. The rise was without 

 break or considerable irregularity. It was : 



Until January 14, 4:18 p. m.. 3.0"^"^ Until January 17, 3:18 P. M„ 128.0"^"* 

 Until January 15, 3:18 P. m., 64.5 Until January 18, 3^18 P. M., iS^o 



Until January 16, 3:18 P. M., I01.5 



To obtain the decrease in tension in this manometer these 





