in 1908. An afternoon in October, 1909, was spent in felling and cutting up a tree cactus (Carnegiea or Cerus giganteus), near the Desert Laboratory, which consisted of a single cylindrical trunk 18 feet in height. The total weight was nearly a ton, and a section was found to contain over 91 per cent, of water, showing that the entire plant held over seventeen hundred pounds of water, or about five barrels.
It has previously been pointed out that during the dry season these plants sustain only an anchorage relation with the soil, and that absorption ceases wholly. The experiments were therefore planned to detach a number of individuals of the sahuaro (Carnegiea), the melon cactus (Echinocactus), and various opuntias from the soil, place them on suitable supports in the accustomed upright position and thus simply lengthen the dry seasons to which they had been subject. Accidents in nature tear man) individuals loose from the soil and they may not be able to perfect a new root-system for many months, so that the observations closely simulated happenings in the history of the species involved. Some of the test plants were placed in the open air, some in the more equable conditions of a well lighted laboratory room, a few were kept for periods of a few months in constant temperature dark room, and others were exposed to the full blaze of the Arizona sunlight, standing on a base of black volcanic rock, thereby avoiding none of the desiccating effects of the climate.
The formidable armature of the bulky bodies of these plants made their manipulation a matter of some difficulty and discomfort even with the best supports and harness that could be devised. The larger ones were placed on platform scales, where they were allowed to remain undisturbed. The majority, however, were mounted and two or three men were necessary to handle them in the weighings which were made at intervals correlated with the season and the rate of loss.
All individuals showed a high rate of loss when first taken from the soil, the excess being attributed to the evaporation from abraded surfaces of the roots and stems. Next it was found, of course, that the rate of loss was least during the cooler season, at which time an Echinocactus might lose as little as one forty-thousandth of its weight in a day, and on the other hand during many days in the hot dry season the daily loss was one three-hundredth of its total weight. The minimum of the tree cactus was one nine-thousandth of its total and the maximum was about that of the Echinocactus, although not measured under equivalent conditions.
Chief interest in the rate of loss, however, centers about the behavior of these plants from season to season, especially when the amount of water on hand was taken into account. In work of this kind it is found convenient to use a standard of succulency which calculates the number of c.c. of water to 100 sq. cm. of surface. Thus a great melon cactus