GROWTH IN TREES. 23 



drograph was 9 mm. and the. thickness of this layer (laid down in 1919) 

 taken from a boring on another diameter, in January 1921, was about 

 8 mm. 



Heavy rains beginning on November 26, 1919, were followed by an 

 increase beginning on November 29, which continued until De- 

 cember 14. A similar slight increase following rain took place early 

 in January and late in January. 



The instrument was reset on February 9, 1920, and a second one 

 reached by a scaffolding attached at a point 8 meters above it, both 

 making a record in which the variations in diameter were amplified 

 8 times (see fig. 4). The circumference of the trunk at the upper in- 

 strument was 100 cm. as compared with 130 cm. at the lower one. 

 The levers of both instruments were set to give an amplification of 

 8 times, so that it was possible to make direct comparisons of the ac- 

 tion of the trunk in the two places. The bark in both cases was thinned 

 so that the contacts were made on layers of cork not more than a 

 millimeter in thickness. The bark at the base of the tree was heavy 

 and was coming away in flakes, while that at the upper instrument 

 had begun to show only shallow rifts and was not more than 3 or 

 4 mm. thick at any place. Thermometers thrust under the bark on 

 the north side of the trunk showed temperatures as low as 6 C. on the 

 date mentioned and as low as 8 C. on March 30. 



Some enlargement of the same type followed rain March 18 to 22, 

 but this was followed by a period of reversing variations, and not until 

 April 7 did a continuing enlargement begin which showed an initial 

 rate slightly greater than that at the uppermost instrument, which 

 began on the same day (fig. 5). 



As this manuscript is being completed, the records for the early 

 part of 1921 are available. They show that the tips of the branches 

 on young and old trees began to grow about mid-January, and that the 

 enlargement of the upper part of the trunk of this tree showed en- 

 largement on the 27th of January, 2 days earlier than any increase 

 of the lower part of the trunk. Action in the two places was syn- 

 chronous in 1920, but did not begin until 2 months later than the 

 awakening of the tips of the branches. No attempt has yet been 

 made to correlate these facts with the activities of the roots, but the 

 observations of W. B. McDougall are to the effect that growth of the 

 roots begins quite early in the season and that it depends largely upon 

 moisture and temperature. 1 



As in the previous year, a period of reversible variations began late 

 in May 1920, which continued until July 2, at which time the soil 

 moisture content had come down to 5 to 7 per cent. The total accre- 

 tion at the base of the tree was slightly less than 3 mm., slightly 



1 McDougall, W. B. The growth of forest tree roots. Amer. Jour. Bot., 3:385-392. 1916. 



