this type of instrument. The older and the improved lever sets are illustrated in Plate I, Figs. I and 2.

A dendrometer of simple design has been perfected which may be placed around the trunk of a tree and the size of the trunk read on a dial from time to time. The essential parts of this instrument are an encircling wire engaged with a number of bearing levers. One end of the wire is anchored and the other is attached to the short end of a lever, the free end of which moves over a scale giving readings of the size of the trunk in terms of several radii, or of the circumference.

By the assistance of collaborators measurements of a number of evergreen and deciduous trees in various habitats from the Atlantic seaboard to the Pacific coast were made in 1919 and 1920. Beech, ash, walnut, sycamore, pines, spruce, fir, poplar and oak trees were included in the list.

The principal generalizations supported by the information obtained may be briefly summarized as follows:

The period in which enlargement of trunks takes place is comparatively brief even in places in which the season is of indeterminate duration. Growth is an activity of an embryogenic tract of tissue, the activity of which depends upon environmental conditions, and no part of the observations suggested a seasonal rhythmic action. The Chihuahua pine which exhibits growth of the trunk with that of the branches on the dry mountain slopes with the advance of the temperatures in May and June, is brought to rest coincident with the desiccation of the soil in the dry fore-summer. Reawakening ensues consequent upon the summer rains and enlargement continues until checked by the decreasing temperatures and increased soil desiccation in the autumn.

The Monterey pine (Pinus radiata) shows beginning growth of the trunks with the advance of temperatures January to April, and comes to rest in July with the desiccation of the soil. Quercus agrifolia in the same region begins earlier and ceases to grow in June or July. Both may be awakened in July or August by deep irrigation of the soil (Fig. 3).

The trunks of all the trees measured show a daily variation in size, by which the maximum is reached shortly after sunrise and

the minimum at a time after noon, dependent upon external agencies. These variations appear to depend upon the water-balance in the woody cylinder, are greatest in the seasons in which water-loss

FIG. 3. A, dendrographic record of variations in diameter of Monterey pine tree I meter from the ground for week beginning April 5, 1920. B, record from instrument attached to the trunk 9 meters from the ground. Daily equalizing variations with actual enlargement beginning mid-week X 8, on a scale of 10 mm. intervals.

from the crown is greatest, are least in the cooler or damper seasons, and are to be detected in the records even in the period of most rapid enlargement of the trunk (Fig. 4).

FEB 1-10-1920

6PM 6AM 6PM 6AM 6PM 6AM 6PM 6AM 6PM 6AM 6PM 6AM 6PM 6AM

70

60

50

40

30

20
10

FIG. 4. Auxographic record of elongation of terminal internode of young pine tree showing cessation of growth and shrinkage during the midday period, coincident with the decrease in diameter of trunks of larger trees, X 3, on a scale of 10 mm. intervals.

The trunk of a tree may in fact be compared to the supply hose of a fire engine coupled to a hydrant. When the pressure from the mains is enough to supply water faster than it can be pumped out the hose is distended. When the engine tends to take water faster than it would be delivered by the system, the hose would tend to

collapse. Something of this sort takes place in many trees which have been kept under observation. The conduit in this case however is not a simple pipe or a set of pipes, but is made up of vessels through which water may pass under capillary conditions, and enclosed box-like tracheids which may be only partially filled with water. When water is withdrawn from such a system faster than it is taken in the resulting changes in form and size are complex in character, but are expressed by the well-defined daily equalizing variations which are of a characteristic type for each kind of tree (Fig. 5).

JULY 5-12-1920

6AM 12 6 PM 12 6 AM 12 6PM 12.6 AM 12 6 PM 12 6 AM 12 6PM 12 6 AM 12 6 PM 12 6 AM 12 6 PM 12 6 AM 12 6PM 12 6 AM 12

70

60

50

40

30

20 10

FIG. 5. Dendrographic record of Arizona ash near the end of the growing season. The daily equalizing variation is large, but the diameter shows an increase on each successive day. Variations X 10 on a scale of 10 mm. intervals.

The greatest daily equalizing variations were shown by Fraxinus, Pinus, Picea, Pseudotsuga and Juglans, and lesser variations were displayed by Populus, Platanus, Fagus, Quercus and Citrus. available facts furnish the basis of an adequate explanation of such differences.

Estimates of the range of daily equalizing variations in a Monterey pine taken from bearings on a thin layer of cork external to the bast of a trunk which had ceased to grow for the season shows that the diameter might vary one part in 1,750. That a large share of this variation is due to changes in the hydration of the living cells is proved by the fact that when bearings are taken on the woody cylinder of the trunk internal to the growing layer the variation drops to one part in 8,750 of the diameter. The actual change in volume in the first instance calculated on the basis of a conical trunk 18 meters high and 35 cm. in diameter at the base would amount to

about 400 cu. cm. of which not more than one-fifth or 80 cu. cm. is attributable to variations in the wood. It is to be noted however that the change in the volume of the wood may by no means be taken to represent the water deficit in the wood. The woody mass is made up of box-like cells, which may include a bubble of gas, the water forming no more than a thin film on the wall of the cell and enclosing the gas bubble in the condition of extreme water deficit. The withdrawal of water through the walls of the cells which are semirigid, increases the surface tension of the gas bubble, which results in a slight lessening of volume of the whole mass, but in an amount that would constitute no more than a small fraction of the total of the water loss.

Awakening and growth of the terminal buds with resultant elongation of leaders and branches generally begins some time before enlargement of the trunk takes place in many trees. The period separating the two may be no more than a week in Quercus agrifolia and has been seen to be as much as ten or twelve weeks in Pinus radiata. Observations on the Parry spruce and Douglas fir show that the trunks of these trees are enlarging at a time when the buds are in a very early stage of enlargement.

In the single case in which dendrographs were attached to a pine tree I meter and 8 meters above the ground growth began coincidently at the two places in 1920. In the following year, however, the dendrograph at the higher point on the trunk recorded enlargement a few days before any action near the ground was made visible. In February, 1921, an auxograph was brought into bearing on the internode of a pine tree five or six years old which had been formed in 1919. The buds had made a growth of 4 or 5 cm. but no action had yet begun in the internode. A second instrument was brought into bearing on the middle of the internode formed in 1920 on another young tree. Steady enlargement was in progress.

The embryonic layer of a tree is in the form of an enclosing sheath terminating in the cones of the growing points. Activation of this tract is generally initiated in the growing points. Swelling in the cambium layer may be practically coincident with this awakening in some trees. Cases are recorded in the following paper in which weeks elapsed between the awakening of the buds and the

enlargement of the base of the trunk. Activation of the growing cells may be taken to depend upon the localized food-supply, temperature, moisture or other factors.

The fact that growth depends upon physical conditions largely external instead of being a manifestation of a rhythm on the part of the tree is well evidenced by tests in which trees which had ceased to grow with the seasonal drying out of the soil were awakened by a renewed water supply.

Irrigation of the soil which had a moisture content of less than 6 per cent. around the roots of a Monterey pine was followed by progressive enlargement constituting growth at the base of the tree, and at a point 8 meters higher within 24 hours. The distance from the absorbing surfaces of the roots through which the added water supply must enter could not be less than 3 meters from the lower instrument and the influence of the added supply was felt at the upper instrument II meters from the absorbing surfaces within the day. It does not seem possible that water could have been conducted. through the tracheids this distance within the given length of time.

An irrigation test similar to the above was made with a small California live oak (Quercus agrifolia). The results were even more startling than those described for the pine. Within two hours. the dendrograph which had its contacts with the tree at least 3 meters from the absorbing surfaces showed some enlargement, an action which may be directly connected with the fact that the vessels in this oak are numerous and large.

The irrigation experiments might be held to simulate the effects. of stream overflow, which if due to melting snows would not be accompanied by any marked higher humidity. It is seen to result in the formation of a tapering shell of wood which was as thick as the seasonal formation at the base of the trunk, but which had but half this thickness 8 meters higher up on the trunk. The layer of normal formation was of practically identical thickness at the two places.

One of the earliest series of measurements which would enable the forester to follow the seasonal variations of trunks was that made by C. E. Hall at San Jorge, Uruguay, in 1885-1890, by commonly used methods of calibration.3 Friedrich perfected a device for

3 Hall, C. E., "Notes on Tree Measurements Made Monthly at San