Fig. 59.  Vicia faba:  germinating seeds, suspended in damp air:  A, with radicle growing perpendicularly downwards; B, the same bean after 24 hours and after the radicle has curved itself; r. radicle; h, short hypocotyl; e, epicotyl appearing as a knob in A and as an arch in B; p, petiole of the cotyledon, the latter enclosed within the seed-coats.

of 23 germinating seeds, which were pinned to the lids of jars, generally with the hilum downwards, and with their radicles pointing to the centre of the earth.  After an interval of from 24 to 48 hours the black lines on the hypocotyls of 16 out of the 23 seedlings became distinctly curved, but in very various degrees (namely, with radii between 20 and [page 93] 80 mm. on Sachs’ cyclometer) in the same relative direction as shown at B in Fig. 59.  As geotropism will obviously tend to check this curvature, seven seeds were allowed to germinate with proper precautions for their growth in a klinostat,* by which means geotropism was eliminated.  The position of the hypocotyls was observed during four successive days, and they continued to bend towards the hilum and lower surface of the seed.  On the fourth day they were deflected by an average angle of 63o from a line perpendicular to the lower surface, and were therefore considerably more curved than the hypocotyl and radicle in the bean at B (Fig. 59), though in the same relative direction.

It will, we presume, be admitted that all leguminous plants with hypogean cotyledons are descended from forms which once raised their cotyledons above the ground in the ordinary manner; and in doing so, it is certain that their hypocotyls would have been abruptly arched, as in the case of every other dicotyledonous plant.  This is especially clear in the case of Phaseolus, for out of five species, the seedlings of which we observed, namely, P. multiflorus, caracalla, vulgaris, Hernandesii and Roxburghii (inhabitants of the Old and New Worlds), the three last-named species have well-developed hypocotyls which break through the ground as arches.  Now, if we imagine a seedling of the common bean or of P. multiflorus, to behave as its progenitors once did, the hypocotyl (h, Fig. 59), in whatever position the seed may have been buried, would become so much arched that the upper part would be doubled down parallel to the lower part; and

* An instrument devised by Sachs, consisting essentially of a slowly revolving horizontal axis, on which the plant under observation is supported:  see ‘Würzburg Arbeiten,’ 1879, p. 209. [page 94]

this is exactly the kind of curvature which actually occurs in these two plants, though to a much less degree.  Therefore we can hardly doubt that their short hypocotyls have retained by inheritance a tendency to curve themselves in the same manner as they did at a former period, when this movement was highly important to them for breaking through the ground, though now rendered useless by the cotyledons being hypogean.  Rudimentary structures are in most cases highly variable, and we might expect that rudimentary or obsolete actions would be equally so; and Sachs’ curvature varies extremely in amount, and sometimes altogether fails.  This is the sole instance known to us of the inheritance, though in a feeble degree, of movements which have become superfluous from changes which the species has undergone.