the latter came into play; and they all eight became bowed downwards to the centre of the earth in 20 h., excepting one which was only slightly acted on.  Two of them were a little bowed downwards in only 5 h.!  Therefore the cards, affixed to the lower sides of their tips, seemed to produce no effect; and geotropism easily conquered the effects of the irritation thus caused.  Secondly, 5 oldish radicles, 1 ½ inch in length, and therefore less sensitive than the above-mentioned young ones, were similarly placed and similarly treated.  From what has been seen on many other occasions, it may be safely inferred that if they had been suspended vertically they would have bent away from the cards; and if they had been extended horizontally, without cards attached to them, they would have quickly bent vertically downwards through geotropism; but the result was that two of these radicles were still horizontal after 23 h.; two were curved only slightly, and the fifth as much as 40o beneath the horizon.  Thirdly, 5 beans were fastened [page 153] with their flat surfaces parallel to the cork-lid, so that Sachs’ curvature would not tend to make the horizontally extended radicles turn either upwards or downwards, and little squares of card were affixed as before, to the lower sides of their tips.  The result was that all five radicles were bent down, or towards the centre of the earth, after only 8 h. 20 m.  At the same time and within the same jars, 3 radicles of the same age, with squares affixed to one side, were suspended vertically; and after 8 h. 20 m. they were considerably deflected from the cards, and therefore curved upwards in opposition to geotropism.  In these latter cases the irritation from the squares had over-powered geotropism; whilst in the former cases, in which the radicles were extended horizontally, geotropism had overpowered the irritation.  Thus within the same jars, some of the radicles were curving upwards and others downwards at the same time—­these opposite movements depending on whether the radicles, when the squares were first attached to them, projected vertically down, or were extended horizontally.  This difference in their behaviour seems at first inexplicable, but can, we believe, be simply explained by the difference between the initial power of the two forces under the above circumstances, combined with the well-known principle of the after-effects of a stimulus.  When a young and sensitive radicle is extended horizontally, with a square attached to the lower side of the tip, geotropism acts on it at right angles, and, as we have seen, is then evidently more efficient than the irritation from the square; and the power of geotropism will be strengthened at each successive period by its previous action—­that is, by its after-effects.  On the other hand, when a square is affixed to a vertically dependent radicle, and the apex begins to [page 154] curve upwards, this movement will be opposed by geotropism acting only at a very oblique angle, and the irritation from the card will be strengthened by its previous action.  We may therefore conclude that the initial power of an irritant on the apex of the radicle of the bean, is less than that of geotropism when acting at right angles, but greater than that of geotropism when acting obliquely on it.