In the two experiments just given, the cotyledons moved backwards or from the window shortly after being subjected to darkness; and whilst tracing the circumnutation of various kinds of seedlings exposed to a lateral light, we repeatedly observed that late in the evening, as the light waned, they moved from it.  This fact is shown in some of the diagrams given in the last chapter.  We wished therefore to learn whether this was wholly due to apogeotropism, or whether an organ after bending towards the light tended from any other cause to bend from it, as soon as the light failed.  Accordingly, two pots of seedling Phalaris and one pot of seedling Brassica were exposed for 8 h. before a paraffin lamp, by which time the cotyledons of the former and the hypocotyls of the latter were bent rectangularly towards the light.  The pots were now quickly laid horizontally, so that the upper parts of the cotyledons and of the hypocotyls of 9 seedlings projected vertically upwards, as proved by a plumb-line.  In this position they could not be acted on by apogeotropism, and if they possessed any tendency to straighten themselves or to bend in opposition to their former heliotropic curvature, this would be exhibited, for it would be opposed at first very slightly by apogeotropism.  They were kept in the dark for 4 h., during which time they were twice looked at; but no uniform bending in opposition to their former heliotropic curvature could be detected.  We have said uniform bending, because they circumnutated in their new position, and after 2 h. were inclined in different directions (between 4o and 11o) from the perpendicular.  Their directions were also changed after two additional hours, and again on the following morning.  We may [page 465] therefore conclude that the bending back of plants from a light, when this becomes obscure or is extinguished, is wholly due to apogeotropism.*

In our various experiments we were often struck with the accuracy with which seedlings pointed to a light although of small size.  To test this, many seedlings of Phalaris, which had germinated in darkness in a very narrow box several feet in length, were placed in a darkened room near to and in front of a lamp having a small cylindrical wick.  The cotyledons at the two ends and in the central part of the box, would therefore have to bend in widely different directions in order to point to the light.  After they had become rectangularly bent, a long white thread was stretched by two persons, close over and parallel, first to one and then to another cotyledon; and the thread was found in almost every case actually to intersect the small circular wick of the now extinguished lamp.  The deviation from accuracy never exceeded, as far as we could judge, a degree or two.  This extreme accuracy seems at first surprising, but is not really so, for an upright cylindrical stem, whatever its position may be with respect to the light, would have exactly half its circumference illuminated and half in shadow; and as the difference in illumination of the two sides is the exciting cause of heliotropism, a cylinder would naturally bend with much accuracy towards the light.  The cotyledons, however, of Phalaris are not cylindrical, but oval in section; and the longer axis was to the shorter axis (in the one which was measured) as 100 to 70.  Nevertheless, no difference could be