_a_, Mature, twisting beginning; _b_, separate fruit, fully twisted.]
lower half, which when dry is much twisted, like a rope (Fig. 11). The covering of the seed itself is furnished with stiff short hairs pointing upwards. The whole structure when mature is cast off by the parent. The curiously twisted appendage is hygroscopic, and readily responds to wetness by untwisting and to dryness by twisting. Should it be thus caused to untwist when the upper end is free from obstruction the latter will revolve slowly like the hand of a clock. But should it meet with an obstacle in the course of its revolutions, such as a blade of grass, the motion is transferred to the lower end, which revolves like an auger, and, lengthening as it untwists, forces the seed into the ground. Should dryness supervene, the backward-pointing hairs on the seed-envelope prevent its being drawn out again when retwisting and consequent shortening take place. These _Erodium_ fruits are among the most interesting in the British flora, and are well worth experimenting with.
2. _Water._--Water, which forms the most frequent and the most serious barrier to plant migration, under certain circumstances is a very efficient agent of dispersal. At the same time, its powers in the latter direction are strictly circumscribed. As regards fresh water, seeds which float may be wafted across lakes. Rivers are more effectual, as seeds may be transported long distances in their currents and thrown up finally on their banks or over flooded areas. When we consider the sea, we realize that there is here a possibility of almost unlimited dispersal provided that the seeds are not injured by salt water, and that they can remain afloat. It is on the latter point that the whole efficacy of water dispersal turns. This was long ago recognized, and investigations have been made by many naturalists to determine the buoyancy of seeds of all kinds. The results show that, taking the seeds of the plants of any country as a whole, not more than about 10 per cent. are capable of floating for more than a short period, while most of them sink at once in either fresh or salt water. So one’s vision of seeds transported in myriads over hundreds of miles of sea is rudely dispelled; and the fact that many seeds can survive prolonged immersion in sea-water uninjured is of little account. The 10 per cent. of our own flora which produce buoyant seeds are mainly riverside and seaside plants; and no doubt their dispersal is to a great extent due to streams and tidal currents. But the majority of the hundreds of thousands of seeds which a river transports annually find their last resting-place in quiet backwaters or on the floor of the sea.