16 GROWTH IN TREES. 
agrifolia, 75 cm. in circumference, on May 28, 1920. Spliced banjo 
wires of silvered steel were run through the outer ends of these plungers 
or contact rods, one end being anchored in such manner that slack 
could be taken up and tension adjusted ; the other end was taken around 
a small pulley, one radius of which was prolonged by a pointer 15 cm. 
long. The radius of the pulley was 0.5 mm. Five contact points 
were arranged on the tree which would allow this instrument to show 
variations in as many radii. Six months later the tip of the pointer 
had moved through 110 cm. and as the amplification was 30, the actual 
enlargement of the perimeter of the polygon made by the wire was 
* em., which was due to the action of the tree in 5 radii. This elon- 
gation being of the circumference, the average elongation of a radius 
would be represented by 
i's prewe 1 
feXE X a74i6 = ™ nearly 0.5 mm. 
This would imply an enlargement of a millimeter in thickness. It 
is to be noted that the observation was begun at a time when the tree 
had made the greater part of its growth for the season. 
After a series of trials of a wide variety of devices, a final design was 
adopted in which the support was a narrow band of galvanized iron 
or a belt of heavy wire bent in undulating folds, the ends being securely 
brought together by slender threaded rods working in slotted angle- 
irons attached to the ends of the belt. Sections of copper tubing or 
copper rods, 15 cm., long, are bent into the form of an L. One arm, 
10 cm. long, is flattended and perforated for attachment to the belt or 
support. The other arm is in a position radial to the tree and the 
spring-like action of the long arm presses its end gently against a pre- 
pared spot in the bark. A pointer and dial are attached to the support 
at the point where the ends are brought together. One end of a piece 
of nickel wire, about 28 or 30 gage, is anchored to a post on the dial, 
and the wire is led around the tree, passing through suitably smoothed 
holes in the short arms of the copper arms or plungers, and the free 
end is attached to the short arm of the pointer (fig. 3). 
Adjustment of either end of the wire is made so that a delicate but 
positive pressure is exerted by each plunger on the surface of the trunk 
and the pointer is at zero, or at some datum point duly noted. Such 
instruments are inexpensive, so that they might be used in large num- 
bers in any serious attempt to measure increments of timber trees, and 
would need very little attention. It is to be noted, of course, that they 
may be subject to accidents from falling branches or large animals; or, 
