704 
MR. G-. J. ROMANES ON THE LOCOMOTOR SYSTEM OF MEDUSAE. 
a stimulus applied, not only as before at f, but also at e or d. In such comparative ex- 
periments, however, it is to be understood that the higher up in the bell a cut is placed, 
the shorter it must be ; for a fair comparison requires that the two ends of the cut 
shall always touch the same two radii of the nectocalyx. Still, if the cut is only a very 
short one (say 1 or 2 millims. long), this consideration need not practically be taken 
into account ; for such a cut, if situated just above the margin, as represented at a , will 
have the effect of destroying the localizing power of the polypite as regards the 
corresponding arc of the margin ; but if situated high up in the bell, even though its 
length be still the same, it will not have this effect. 
From all this, then, we have seen that the connexions which render possible the 
accurate localizing functions of the polypite are almost, though not quite, exclusively 
radial. We have also seen that between accurate localization and mere random move- 
ments on the part of the polypite there are numerous gradations — the degree of decline 
from one to the other depending on the topographical relations between the point of 
stimulation and the end of the section (the section being of the form represented by 
AB in the figure). These relations, as we have seen, are the more favourable to correct 
localization, (a) the greater the radial distance between the point of stimulation and 
the end of the section ; and (b) the less the circumferential distance between the point 
of stimulation and the radius let fall from the end of the section. But we have seen 
that the limits, as regards severity of section, within which these gradations of localizing 
ability occur are exceedingly restricted — a cut of only a few millims. in length, even 
though situated at the greatest radial distance possible, being sufficient to destroy all 
localizing power of the polypite as regards the middle point of the corresponding arc 
of the margin, and a stimulus applied only a few millims. from the attached end of a 
severed marginal strip entirely failing to cause localizing action of the polypite. Lastly, 
we have seen that even after all localizing action of the polypite has been completely 
destroyed by section of the kinds described, this organ nevertheless continues actively, 
though ineffectually, to search for the seat of irritation. 
(d) The last-mentioned fact shows that after excitational continuity of a higher order 
has been destroyed, excitational continuity of a lower order nevertheless persists. Or, 
to state the case in other words, the fact in question shows that after severance of the 
almost exclusively radial connexions between the bell and the polypite by which the 
perfect or impaired localizing function of the latter is rendered possible, other connexions 
between these organs remain which are not in any wise radial. I therefore next tested 
the degree in which these non-radial connexions might be cut without causing 
destruction of that excitational continuity of a lower order which it is their function 
to maintain. It will here suffice to record one mode of section which has yielded 
important results. A glance at the accompanying figure (8) will show the manner 
in which the Medusa is prepared. The margin having been removed (in order to pre- 
vent possible conduction by the marginal nerve-fibre), a single deep yadial cut (a a) 
is first made, and then a circumferential cut ( a , b, c ) is carried nearly all the way 
