placed under a smooth straight root, but in twenty-four hours the root had curved and grown past the surface, a tuft of hairs on the curve indicating that some retardation of growth took place. For over a week the curling and hair production continued, then the root grew horizontally and struck the glass side. It became kinky and hairs continued to be formed for five days.

FIG. 11.-Diagram of apparatus to stop the growth of a root by a glass rod.

The fact that kinking takes place in free water shows that some other factor or factors besides resistance must be acting, but the facts brought out in the two experiments make it appear possible that resistance may be a partial cause for the kinking and hairiness of roots.

As the plaster cup in the preceding experiment was unsuccessful, glass tubes were tried. They were of sufficiently small bore to prevent a relief of pressure by too great bending. Smooth roots of corn were repeatedly allowed to grow into glass tubes (fig. 12). Usually the tip became more or less swollen, nearly or quite filling the tube. Primary roots showed kinking at the bottom, and hairs appeared in diminishing lengths from the bottom to the top. Only a few hairs appeared on the adventitious roots. When kept at high temperatures (24-34°C) the roots grew smooth, although bent and curved. If the resistance were relieved by allowing the roots to curve above the tube, hairs ceased to appear, conforming with the statement of MER (51, p. 584) that feeble retardation is not able to produce hairs. In one experiment the growth appeared to be so great that the roots were crushed and broken, producing no hairs on these portions.

C. Wounding.

SCHWARZ (75, p. 158) was not able to cause hair formation by cutting off the roots 2-10mm from the tip, nor by burning the tip with caustic. In my experiments the results were various according to the conditions. If the wound were not of sufficient depth to retard growth, if it were beyond the elongating zone, or if the plants were grown in warm water, no swelling or hairs appeared; otherwise hairs were produced. Thus in corn seedlings the tips of primary roots were pinched off about 1mm from the tip. One showed hairs upon the swollen tips; another sent out a tuft of hairs and then grew smooth. In the latter case the wound was not of sufficient depth to more than slightly retard the growth. Of roots cut and burned, several showed hairs, the burned ones curving; several simply stopped growing and produced laterals; while others showed no effect. The cut tips of corn roots growing in air and producing short hairs became slightly swollen in twenty-four hours, and long hairs appeared above the cut. Both in light and darkness hairs were produced above the cut, whether the swelling appeared or not. This may have been due to the appearance of new hairs among the old ones, or to the stimulated growth of some of the old hairs, but more probably to the retardation of the zone in process of formation when the operation was performed. DEVAUX (10, p. 308) states that new hairs may appear among the old ones, but appearances which might be interpreted in that manner might be due to arrested development of some of the hairs. This would be difficult to decide, unless hairs were actually seen to originate between others (fig. 2). SCHWARZ (75, p. 165) and HABERLANDT (75, p. 187) state that hairs are always produced in acropetal succession.

D. Medium.

SACHS (71, p. 410) found that roots of land plants grew more rapidly in soil than in air or water, and his results have been confirmed by WACKER (84, pp. 109-115). The latter, however, found that in slimy soil the growth was retarded more than in water, and the denser the material the slower the growth. PFEFFER (66, p. 320) says the rate of growth is not affected by the density of the medium, roots growing as rapidly in fluid clay as in water. These conflicting

results are due probably to the different amounts of water in the soils used.

Roots of corn grown in ground quartz, garden soil, and air gave these results: in quartz, av. length 19.5mm, hairs abundant; in soil, 22.3mm, hairs good; in air, 50mm, hairs poor. From these figures it seems that the resistance of the substratum bears direct relation to hair production; but the factor of water supply has undoubtedly an important influence, the quartz being less compact and therefore drier than the garden soil. Other experiments showed slower growth in air and quartz than in soil or water.

The behavior of roots of Elodea in the substratum has been mentioned, with the suggestion that retardation due to the soil particles was the principal factor. It will be shown later that a diminution of oxygen supply has a tendency to suppress hair production. There is less oxygen in the substratum than in the freely flowing water above it. It appears, therefore, that retardation, whether from soil resistance or chemical influence, must be the chief factor in producing the kinking and the hairs. Whether the hairs are due to the kinking, or both are due to the retardation of growth, cannot be stated. The production of hairs by retarding growth with glass tubes took place at times without kinking, though in the majority of cases the two results were associated. SCHWARZ (75, p. 159) considers “nutation" (kinking) the most potent factor in the production of root hairs, but it seems as if they might both be referred to unequal retardation of the growth of the root. Measurements of the epidermal cells of roots of Elodea give the following averages in millimeters:

Here the soil roots show better hairs than the quartz, and have the shortest cells when hairy. As will be seen later, however, the comparative lengths of cells of different roots can only be taken as supporting not as decisive evidence.

Corn seedlings were allowed to send their roots between glass

plates, on one of which was a layer of paraffin with sections covered with dune sand and ground quartz. The growth over the paraffin was smooth; the roots running over the sand were wavy, in some places producing hairs; and the one on the quartz kinked with more hairs (fig. 13). One root from the plant growing over quartz

FIG. 13.-Corn seedlings growing in water in a glass jar between paraffined glass plates, on which was spread in the center a layer of coarse ground quartz; on the right is dune sand; on the left is clean paraffin.

wandered into the paraffin section, curled, and developed hairs. This appeared at the same time as the curling of the main root and may have been correlated with it. A second experiment with the sections horizontal also showed the laterals wavy at the same time that the main root kinked on the quartz.

WATER CONTENT.

According to several investigators (WIESNER 88, p. 149; PFEFFER 64, p. 100; PALLADIN 59, p. 371; BRENNER 7, p. 435; MACDOUGAL

47, p. 64; et al.) the attenuation of the axial members of etiolated plants, where it occurs, is due, in part at least, to a lack or diminution of transpiration. A greater proportion of water in etiolated plants is reported by MACDOUGAL (47, p. 64), PALLADIN (60a), JUMELLE (28, p. 386), et al.

Whether turgor is the cause or the result of growth, elongation of a cell is directly connected with its turgescence, greater water content producing greater elongation. Besides instances of etiolation, this is shown by the curling of roots which rest on the surface of water (SACHS 71, pp. 398-9); by the rounding up of filamentous algae (LIVINGSTON 43, p. 308; 44, pp. 310-312) and of fungi (RACIBORSKI 68, p. 111) upon withdrawal of water by osmotic solutions, and by the tendency of cells to stretch radially from loss of water by transpiration (KOHL 31а, p. 297). The more turgid a cell becomes, therefore, the greater the tendency to stretch in a longitudinal direction. The water content of the root cells may be affected by changing the moisture content of the air, by altering the water content of the soil, or by surrounding them with solutions of higher osmotic pressure.

A. Transpiration.

It is well known that aerial transpiration favors the production of hairs upon aerial organs, WOLLNY (89, pp. 418-435) reporting an increased number of piliferous cells by count. On the other hand, some hairy-leaved plants grown in an aquatic habitat become smooth (COSTANTIN 8, p. 40), and many have noted the absence of hairs on roots in water. Experiments to determine the effect of transpiration from leaves upon the development of root hairs gave negative results. Roots grown in saturated air at various temperatures showed few or no hairs, and any change that reduced the moisture content favored their development. Control plants showed that the temperatures used could not alone produce the results.

B. Saturated soil.

For these experiments corn and wheat were chosen because the former is very sensitive to the inhibitory effect of water, and wheat readily develops hair in that medium.