The effects of graduating the doses of caffein may be stated in connection with another species, viz. Tiaropsis diademata. In a weak solution the effects are a quickening of the pulsations (_e.g._ from 64 to 120 per minute) together with a lessening of their force. On slightly increasing the dose, the pulsations become languid, and prolonged pauses supervene. If the dose is again somewhat strengthened, the pulsations become weaker and weaker, till they eventually cease altogether. The animal, however, is now in a condition of exalted reflex irritability; for its response to a single stimulus consists not merely, as in the unpoisoned animal, of a single spasm, but also, immediately after this, of a series of convulsive movements somewhat resembling swimming movements destitute of co-ordination. If the strength of the solution be now again increased, a stage of deeper anæsthesiation may be produced, in which the Medusa will only respond to each stimulation by a single spasm. In still stronger solutions, the only response is a single feeble contraction; while in a nearly saturated solution the animal does not respond at all. But even from a saturated solution Tiaropsis diademata will recover when transferred to unpoisoned water. 4. _Strychnia._--The species of covered-eyed Medusa which I shall choose for describing the action of strychnia is Cyanæa capillata, which is most admirably adapted for experiments with this and some of the other alkaloid poisons, from the fact that in water kept at a constant temperature its pulsations are as regular as are those of a heart. After Cyanæa capillata has been allowed to soak for ten minutes or so in a weak sea-water solution of strychnia, unmistakable signs of irregularity in the pulsations supervene. This irregularity then increases more and more, till at length it grows into well-marked convulsions. The convulsions manifest themselves in the form of extreme deviations from the rhythmical contractions so characteristic of Cyanæa capillata. Instead of the heart-like regularity with which systole and diastole follow one another in the unpoisoned animal, we now have periods of violent and prolonged systole resembling tonic spasm; and when the severity of this spasm is for a moment abated, it is generally renewed before the umbrella has had time again to become fully expanded. Moreover, the spasm itself is not of uniform intensity throughout the time it lasts; but while the umbrella is in a continuously contracted state, there are observable a perpetual succession of extremely irregular oscillations in the strength of the contractile influence. It is further a highly interesting fact that the convulsions are very plainly of a _paroxysmal_ nature. After the umbrella has suffered a prolonged period of convulsive movements, it expands to its full dimensions, and in this form remains for some time in a state of absolute quiescence. Presently, however, another paroxysm supervenes,