1. The original axial susceptibility gradient of the alga Griffithsia is more or less completely obliterated or reversed after one or two days of confinement in a small volume of water, and the larger the volume of living plant protoplasm introduced, the more rapid the changes. 2. Where the conditions of confinement are extreme, obliteration or reversal of the gradient is followed in one or two days by death with a basipetal gradient, but under less extreme conditions the alteration of the gradient is followed by the separation of the axis into individual cells or occasionally small cell groups, the separation beginning apically and progressing basipetally. 3. Obliteration or reversal of the original axial gradient, followed by cell separation occurs, though more slowly, when the plants are kept in open standing water and to a greater or less degree, but still more slowly, in slowly running water in diffuse daylight. 4. A variable percentage of the isolated cells dies, the death rate being in general highest in apical cells and decreasing with the level of the cell in the axis, at least to the middle levels, below which there is often little difference. The cells which do not die may undergo a new development, giving rise to new apical cells and so to new axes. 5. Determinations of the susceptibility gradients in these reproductive developments from isolated cells show that the new apical cell arises from the region of highest susceptibility in the old cell. This region is more commonly basal than apical, because the gradient is usually reversed at the time of cell separation, but both ends of the cell may give rise to new apical cells where double opposed gradients exist. Rhizoids, on the other hand, arise from low levels in the axial gradient. 6. These experiments show that the original polarity of the cells of Griffithsia can be completely obliterated or reversed, and that the morphological development of new axes is an expression of gradient relations present in the cell at the time. 7. The facts support the conclusion that a gradient in metabolic rate, protoplasmic condition, or whatever we prefer to call it, of which the susceptibility gradient is within certain limits an indicator, constitutes physiological polarity in protoplasm, and that such a gradient is not an inherent property of protoplasm, but may be determined and altered by external factors.