elaborations of the cuticle or adjacent cells in the form of pits, plugs, or other devices which might be regarded as modifying the effect of the stomatal aperture. We are therefore face to face with the relation of the stomatal openings to the regulation of loss of water, unmodified by other anatomical complications.

The small amount of chlorophyll (or its absence, if this be the case) in the stomata of Verbena, their size and mobility, the relatively small amount of starch present, together with other features, have made this plant most valuable in the pursuit of light upon the internal economy of the stomatal apparatus.

METEOROLOGY.

A more exact appreciation of the precise degree of adaptive value of the stomatal mechanism of these plants may be had if we glance for a moment at the conditions under which they live. There are introduced for this purpose the tables of temperatures for the year 1906 and of rainfall for 1905 and 1906. The latter are plotted in fig. 6, together with a curve of normal rainfall (Coville & MacDougal, 1903).

FIG. 6. Curves of rainfall for 1905 and 1906 at the Desert Botanical Laboratory, together with a curve of normal rainfall, based upon data for 15 years taken at Tucson, Arizona.

With regard to the temperatures it will be noted that extreme cold is not known on Tumamoc Hill, where the plants studied are to be found. Growth is therefore possible for the vegetation of the region at almost any time of the year if the soil-water is sufficient. Thus, the ocotillo began to develop its foliage during the latter half of December, 1906, and by the end of the year many individuals were conspicuous for their abundant green vestiture. During the same period the verbena recommenced its growth and at the close of the month many plants showed a fine development, though of course, owing to the lower general temperatures, the growth was in neither of these

plants rapid as compared to that which ensues after the spring and summer rains. During the cool and moist period of the winter season the conditions approximate closely those of a northern cool, moist, though not wet spring, when the ground is wet but the air cool, a high relative humidity, with dews and with an abundance of sunshine. The leaves of the plants which grow at this time are in no danger of wilting, and indeed never show it, and, from the teleological point of view, there is no reason why the stomata should not be wide open. I have, however, not found them open more than six-tenths their possible extent of opening during the warm period of the day, though exceptional stomata show a somewhat wider opening.

TABLE 7.-Rainfall, Tumamoc Hill, Tucson, Arizona, for 1906, in inches.

The approach of the summer is marked by cessation of rain, increase of temperature, and lowering of the relative humidity.* At the same time,

*No accurate data for relative humidity for a sustained period are available, so that general statements must suffice. With reference to the possible high humidities in a given area in the absence of rain see Cannon, 1906. The problem of the desert in this regard is one which commends itself to study. Many areas have a distinctly desert character, but are favored by high atmospheric humidities, and these are contiguous in many places to the desert areas having the opposite condition, e. g., Southern California and Arizona (McGee, 1906).

the soil-water is depleted (Livingston, 1906), though sufficient may be retained to carry some forms of vegetation through to the summer rainy period. Verbena ciliata is in this respect remarkably hardy and, from its anatomical character, quite unexpectedly so. The height of the dry season is one which is trying in the extreme, one would think, and still, with a supply of water TABLE 8.-Rainfall, Tumamoc Hill, Tucson, Arizona, for 1905, in inches.

TABLE 9.-Temperatures for the year 1906, Tumamoc Hill (°F.).

Rainfall.

0.465

2.24

0.048 4.48

0.90

23.183

TABLE 9.-Temperatures for the year 1905, Tumamɔc Hill (°F.)—Continued.

Min. Max. Min. Max. Min. Max. Min. Max. Min. Max.

not obviously great, the foliage of the ocotillo and of Verbena will continue normal for a fairly extended period. When the soil-water is sufficient there is no evidence in the behavior of the stomata that even the highest temperatures and strongest insolation affect the plants unfavorably. As I shall show, when the conditions are so trying as to bring about wilting, the stomata then close; otherwise they open to their full extent during certain portions of the day, maximum dimensions having been observed day after day during the hottest weather. And, moreover, as will appear, the differences of temperatures are not without relation to the physiological processes evident within the guard-cells, although these are not in the direction of reducing the transpiration rate when, aside from the condition of wilting, the danger from transpiration would appear to be greatest.

METHODS.*

In order to answer the questions which I set before myself, it was necessary to study separately the phenomena of transpiration and of stomatal action. The methods employed will now be described.

TRANSPIRATION.

Inasmuch as it was impossible to use potted plants or to dig up and start complete individuals, the method of weighing, so insistently maintained by some as the only strictly reliable method of measuring the transpiration, was not feasible. The use of such indicators as cobalt chlorid, as used by Stahl and others, while useful for certain kinds of comparative study, was not adapted to the conditions surrounding the experiments, the small leaf area of the ocotillo, the extremely low relative humidity, and the high temperatures all contributing to make it very difficult or impossible to arrange a small closed chamber suitable to the method, which is, moreover, not at all adapted, at least as at present carried out, to the demands of quantitative work. For similar reasons, and others which will be discussed later, the type of hygroscope used by Francis Darwin was not available. Two other methods were tried. One is that inaugurated by MacDougal and elaborated by Cannon (1905 b), in which a Lamprecht polymeter is used to indicate the rise of relative humidity within a closed chamber in which is confined the whole or an attached portion of the plant to be studied. Knowing the capacity of the chamber and the temperature and having taken readings of the hair hygrometer at constant intervals, it is then possible, theoretically at least, to calculate the amount of water-vapor which has escaped into the chamber from the surface of the plant or part inclosed. While the method offers the advantage of allowing transpiration to be studied while the plant remains undamaged and growing in the normal surroundings, the objection may still at this time be raised that the hair hygrometer is an inadequate instrument, as at present constructed. It may, however, be quite possible to construct an instrument of great delicacy and accuracy, and this will probably be accomplished in the future. For the purpose of my work this method proved to be of only limited value, and, while a few of the results have been used, they are considered to be of less value than those obtained in the way now to be described.

The second means at hand was the potometer, an instrument which has been repeatedly criticized on well-known grounds.† For some time the MacDougal potometer, consisting of a horizontal tube of small caliber graduated to 100 mg. was used. A considerable number of experiments were made with this type of potometer, but the form of the whole instrument was at length found to be too cumbersome and unwieldy to lend itself to my purposes. I could not move the instrument about without difficulty, and it was very easy

*The main points were presented before the Botanical Society of America at its meeting in December, 1904.

†Burgerstein (1904) and Clements (1905).