414 TRANSFORMATION OF ENERGY 



the outer and inner walls of one cell (layer 2) ; [h) they occur in different cells ; 

 or (c) both conditions are united. Under all conditions the radial cell-walls have 

 transverse pores and play no part in the twisting. In the fourth layer the pores 

 are laid down in continuous spiral lines directed to the right, running from the 

 front wall, over the lateral walls, to the rear walls ; in other words, we have here 

 to deal not with two intersecting plates but with elements which are spirally 

 arranged. If the wall of such an element be equally capable of swelling, each 

 must, as ZiMMERMANN has shown, twist independently. How a complex of 

 such cells bound into a tissue may bring about a spiral twisting may be easily 

 shown by experiment. Cover a strip of paper thickly with portions of the 

 twisting awn of Stipa, gum these to the paper and to each other, and leave the 

 whole to dry ; the structure will then exhibit a close left-hand twining with the 

 paper external (Steinbrinck, 1888, p. 392). 



This last example illustrates the close relationship which exists between 

 twisting movements and the torsions which we have now to discuss. We 

 have seen that the individual cells must undergo torsion if they are provided 

 with pits arranged in a spiral manner. An entire organ can also undergo torsions 

 similar in all respects to those exhibited by its constituent elements. Darwin 

 (1876) bound together wet awns of Stifa, and found that the bundle underwent 

 torsion when it was allowed to dry, but apparently this principle does not come 

 into play in nature {Anemone'? Eichholz, 1885, 554). Torsions of the whole 

 organ occur much more frequently in consequence of the tendency to twist on 

 the part of the individual elements when these are arranged in concentric layers. 

 Such torsions must arise from a relatively greater extension in the peripheral 

 than in the central region. If we twist a bundle of fibres laid parallel to each 

 other we shall find that each, with the exception of the central ones, describes 

 a spirally curved line, and conversely if twisting of the individual elements occurs 

 there arises torsion in the bundle as a whole. The twisting cells of Stipa, which 

 have the structure of those composing the fourth layer in Erodium, are the 

 chief agents in the movement, and one important condition only must be fulfilled, 

 viz. that the fibres should exhibit an increasing capacity for water absorption 

 longitudinally from without inwards, so that in drying the central ones should 

 contract more than those on the periphery. 



A few words may be said in conclusion on the biological significance of the 

 movements which have been described. Almost all of them have to do with the 

 dispersal of seed. In the great majority of cases fruits burst open or dehisce 

 when dried, part of the fruit wall being thrown off, and the seeds escaping in 

 this way from the capsule. There are many plants, however, such as Anastatica, 

 M esemhryanthemtim, &c., whose fruits close in dry weather and open in wet. 

 Fruits which have the power of ejecting their seeds form more perfect illustrations 

 of these hygroscopic movements ; examples of such are Geranium, the twisting 

 pods of Leguminosae and many others which have not been referred to above, 

 such as Viola, Oxalis, &c. Fruits and seeds which have long twisting awns, 

 such as Erodium, Stipa (and many other grasses), many species of Anemone, 

 &c., are able to force their way into the soil by the torsions which take place 

 in these appendages. 



Let us now turn to the consideration of a series of phenomena which are 

 illustrated by the movements of anthers and sporangia, connected with the ejec- 

 tion of the pollen-grains or spores, and hitherto briefly spoken of as hygroscopic 

 movements. We shall consider first the sporangia of ferns, and especially those 

 of the Polypodiaceae. The sporangium is a stalked, lens-shaped body, en- 

 closing the spores within a multicellular unilamellar wall. Most of the cells 

 of the wall are polyhedral, thin-walled plates, but the edge of the lens is occupied 

 by a ring (annulus) of horseshoe-shaped thick-walled cells, which starts from the 

 stalk and more or less completely encircles the capsule (Fig. 124, a). 



