In view of this anomaly, I was led to think that if the rate of the stimulus-wave is dependent in a large degree on the strength of the stimulus that starts it, the slow rate of nine inches a second might be more than doubled, if, instead of using a stimulus so gentle as not to start a contraction-wave, I used a stimulus sufficiently strong to do this. Accordingly I chose a specimen of Aurelia wherein the occurrence of tentacular waves was very conspicuous, and found, as I had hoped, that every time I stimulated too gently to start a contraction-wave, the tentacular wave travelled only at the rate of nine inches a second; whereas, if I stimulated with greater intensity, I could always observe the tentacular wave coursing an inch or two in front of the contraction-wave. It is remarkable, however, that in this, as in all the other specimens of Aurelia which I experimented upon, the reflex response of the manubrium was equally long, whatever strength of stimulus I applied to the umbrella; or, at any rate, the time was only slightly less when a contraction-wave had passed than when only a tentacular wave had done so. The loss of time, however, appears to take place in the manubrium itself, where the rate of response is astonishingly slow. Thus, if one lobe be irritated, it is usually from four to eight seconds before the other lobes respond. But the time required for such sympathetic response may be even more variable than this--the limits I have observed being as great as from three to ten seconds. In all cases, however, the response, when it does occur, is sudden, as if the distant lobe had then for the first time received the stimulus. Moreover, one lobe--usually one of those adjacent to the lobe directly irritated--responds before the other two, and then a variable time afterwards the latter also respond. This time is, in most cases, comparatively short, the usual limits being from a quarter of a second to two seconds. How much of these enormous intervals is occupied by the period of ganglionic latency, and how much by that of transmission, it is impossible to say; but I have determined that the rate of transmission from the end of a lobe of the manubrium to a lithocyst (deducting a second for the double period of latent stimulation) is the same as the rate of a tentacular wave, viz. nine inches a second. The presumption, therefore, is that the immense lapse of time required for reflex response on the part of the manubrium is required by the lobular ganglia, or whatever element it is that here performs the ganglionic function. _Exhaustion._