If the same experiment is repeated in ordinary air, _i.e._ air not saturated with water, the result is somewhat different. If the twig is suspended vertically, with its _apex upward_, roots soon appear on the basal end that is in the water, but no roots develop above the water. Small protuberances may appear above the water in the places at which roots would develop if the piece were surrounded by a moist atmosphere, but they do not break through the bark. If the piece is then covered by a jar containing air saturated with moisture, these protuberances may become roots. It is clear, therefore, that the dryness of the air has prevented their development. If a similar twig is suspended (in the air) with its _apex downward_, and the lower end in water, root protuberances appear, at first, only around the base, _i.e._ at the upper end. Under the water, at the apical end, small and weak roots may develop, or may even not appear at all. These results agree, in the main, with those in which the piece is surrounded by moist air, and give evidence of an inner polarity that is an important factor in the regeneration. The results show that in a piece with the basal end in water and the rest of the piece in the air the tendency to produce roots above the water is suppressed by the dryness of the air. In an inverted piece, however, with the apex in water, the innate tendency to produce roots at the basal end is strong enough to overcome the effect of the dryness of the air to suppress their development. The abundance of water absorbed by the apex of the piece makes the development of the roots possible under these conditions despite the dryness of the air.[33] There is another factor connected with the submergence of the end of the stem in water that can be shown by putting a longer part of the end under the water. Neither roots, if it is a basal end, nor leaf-buds, if it is an apical end, appear on the deeper parts of the submerged end. This is due, in all probability, to the insufficiency of oxygen in the water, and as a result the buds are prevented from developing. It can be shown that light has also an influence on the regeneration of pieces, and that it has a stronger influence on some plants than on others. In some plants roots develop only on that side of the stem that is less illuminated. In _Lepismium radicans_, for instance, adventitious roots are produced by the plant even in dry air. Pieces of the stem can produce roots on either the upper or the lower surface, according to which side is less illuminated. A piece of the stem of this plant that had been kept in the dark produced two roots, one above and one below,--one, therefore, opposed to the direction of the action of gravity, and the other in the direction of that action. Even in pieces of the willow, suspended in a moist atmosphere, roots develop better and over a greater length of the stem on the less illuminated side.