The dose to be employed is an exceedingly important matter, and one as to which practitioners are still very often injudicious. We ought never to commence with a larger dose than one-sixth of a grain; but very often as little as one-twelfth of a grain will give effective relief, and in not very severe cases it is well worth while to try this smaller quantity. When no larger quantity than one-sixth of a grain is employed we commonly observe no narcotic effects, _i. e._, there is no contraction of pupil, no heavy stupor, and, although the patient very often falls asleep, on waking he does not experience headache, nor is his tongue foul. I cannot too strongly express the opinion that it is advisable by all means to content ourselves with this degree of the action of hypodermic morphia, unless it fails to produce a decided impression on the pain. But in very severe cases our small doses will fail; and then, rather than allow the patient to continue having severe paroxysms unchecked, we must frankly admit the necessity of using a narcotic dose from one-quarter to one-half of a grain, according to circumstances. Whatever actual dose be employed, it is important not to repeat it with unnecessary frequency; once a day in the milder, and twice a day in the more severe cases, will be all that is advisable, save in very exceptional cases: the point being to administer it as quickly as possible after the commencement of an exacerbation. If by these means we can prevent the patient having any severe pains during a period of several days, we often give time to the affected nerve to recover itself so completely, especially with the aid of other measures to be presently mentioned, that the tendency to neuralgia is completely broken through, and we can drop the injections, either at once or by rapid diminution of the dose, and thereafter treat the case merely with tonics, and with the precautionary measures to be dwelt upon under the heading of Prophylaxis. But, if we have been driven to the use of distinctly narcotic doses, and these do not very speedily break the chain of neuralgic recurrence, it will not do to continue to rely upon hypodermic morphia; it will be best to try some of the local remedies (blistering, galvanism) with it. If this combination fails, we should then try the effect of atropine, the sulphate of which, hypodermically injected, fully represents for all useful purposes the mydriatic class of narcotics.