[Illustration: Fig. 6.] [Illustration: Fig. 7.] I may as well state here that in water at all temperatures, within the limits where responses to stimuli are given at all, the staircase action admits of being equally well produced; but in order to procure the _maximum_ effect for any given temperature, the rate at which the successive stimuli are thrown in must be quicker in warm than in cold water. CHAPTER IV. EXPERIMENTS IN SECTION OF COVERED-EYED MEDUSÆ. _Amount of Section which the Neuro-muscular Tissues of the Medusæ will endure without suffering Loss of their Physiological Continuity._ The extent to which the neuro-muscular tissues of the Medusæ may be mutilated without undergoing destruction of their physiological continuity is in the highest degree astonishing. For instance, to begin with the covered-eyed Medusæ, I shall briefly state three modes of section, the results of which serve to show in a striking manner the fact in question. [Illustration: Fig. 8.] [Illustration: Fig. 9.] The annexed woodcuts represent the umbrella of Aurelia aurita, with its manubrium cut off at the base, and the under or concave surface of the umbrella exposed to view, shewing in the centre the ovaries, and radiating from them the branched system of nutrient tubes. The umbrella when fully expanded, as here represented, is about the size of a soup plate, and, as previously stated, all the marginal ganglia are aggregated in the eight marginal bodies or lithocysts. Therefore if the reader will imagine the first of the diagrams (Fig. 8) to be overspread with a disc of muslin, the fibres and mesh of which are finer than those of the finest and closest cobweb, and if he will imagine the mesh of these fibres to start from these marginal ganglia, he will gain a tolerably correct idea of the lowest nervous system in the animal kingdom. Now, suppose that seven of these eight ganglia are cut out, the remaining one then continues to supply its rhythmical discharges to the muscular sheet of the bell, the result being, at each discharge, two contraction waves, which start at the same instant, one on each side of the ganglion, and which then course with equal rapidity in opposite directions, and so meet at the point of the disc which is opposite to the ganglion. Suppose, now, a number of radial cuts are made in the disc according to such a plan as this (Fig. 9), wherein every radial cut deeply overlaps those on either side of it. The contraction waves which now originate from the ganglion must either become blocked and cease to pass round the disc, or they must zigzag round and round the tops of these overlapping cuts. Now, remembering that the passage of these contraction waves is presumably dependent on the nervous network progressively distributing the ganglionic impulse to the muscular fibres, surely we should expect that two or three overlapping cuts, by completely severing all the nerve-fibres lying between them, ought to destroy