cbamber tenuiiiiitin^ in an t-xpiinsioii structure whieli may 

 dilate or contract with the alterations in volume of the Hiiid. 

 It now remains to attach devices for 2)roi)erly demonstrating 

 the raovementH of the upjjer wall of the expansion chamber in 

 order to obtain the chanj^'es induced by variations in tempera- 

 ture. To do this an ii])ri^'lit ))ostis fastened to the up])er wall 

 of the exi)ansi(m chamber and this is provide<l with a sleeve 

 and set screw which allows adjustments of its length in con- 

 nection with a lever (tig. 1, E and F) 4 centimeters long run- 

 ning to an axis held in bearings between two upright stand- 

 ards. A second arm 17 centimeters in length, bearing a suitable 

 pen is attached to this axis and extends so as to come in contact 

 with a suitable uju'ight recording cylinder revolved by clock- 

 work. The long lever is built up of three parts and is provided 

 with a device for adjustment and calibration (tig. 1 H). The 

 cylinder carries double ruled paper and makes a single revo- 

 lution in a week. In the construction of the working model 

 described here, it was found convenient to use the levers and 

 recoi'ding mechanism of a thermograjihof German manufacture, 

 the slips of which were ruled for degrees centigrade. The 

 instrument was calibrated to meet these requirements, but new 

 models now under construction are designed on the Fahren- 

 heit scale. The improved model may also be constructed to 

 carry thermal apparatus for taking air temperatures on one 

 end of the base and the soil temperatures on the other end, 

 thus making synchronous records. 



The peculiar value of the above apparatus consisted in the 

 fact that the recorder could be kept iu an instrument shelter 

 of any convenient pattern and the sensitive bulb carried out 

 and buried in the soil at any desired distance from the shelter, 

 or at any depth in the soil, by which the actual temperatures 

 encountered by the roots of jDlants might be obtained. In 

 practise it was found that the amount of liquid in the capillary 

 tube and ex2)ansion chamber was so small in proi^ortion to that 

 in the bulb that the temperatures encountered by the tube or 

 chamber made no observable error in the reading of the instru- 

 ment. This was true to such an extent that the Hame of an 

 alcohol lamp applied to the tubes did not produce any change 

 in the position of the pen upon the recording slip. The same 

 is true of the etTects of direct exposure to the sun's rays. 



The location for the installation of the instrument for the 

 trial series of records for the first year was a small area of 

 clayey soil mixed with loam lying near the propagating houses 

 and nurseries in the New York Botariical Garden.- The drain- 

 age is free, as the surface of the ground is slightly sloi)ing 

 to the eastward. During the summer the place is covered by 

 the shade of some cherry trees after 4 p. m. 



In the installation of the instrument the recorder, together 

 with an air thermograph, was placed in an instrument shelter 

 of the United States "Weather Bureau pattern. The capillary 

 tube passed through the floor of the shelter to the surface 



