convulsions attending upon diseases or lesions of the nervous centres, of the other organs of the body, and of the blood.

Regarding a chronic paroxysmal affection as one which, in some form or another, must exist during the intervals of attack, and of which the latter is but the occasional although characteristic expression; it is of importance to examine minutely the condition of the patient during the interparoxysmal period, in order to discover whether the morbid state, which underlies the convulsion, and is capable of developing that formidable series of phenomena, has other features than those of the paroxysm. The epileptics, therefore, which have been under my care, I have examined, during the intervals of their attacks, in regard of their mental state; their sensorial and emotional condition; their state in so far as motility is concerned; and also their condition in respect of general health.

The features of the attacks are then described in their several forms; and, under the head of Natural History, I have given information upon such points as the prevalence of epilepsy, its proportion to diseases generally, to affections of the nervous system, and to convulsive diseases. I have examined the causes of epilepsy, both the unavoidable and accidental; and have traced the relations between its several phenomena. By the latter process I have endeavoured to show in what way, and how far changes of mind, of motility, and of the general health, are related to one another; to the duration of the disease; to the nature of its cause; and to the frequency, severity, and form of the attacks. The consequences and complications of epilepsy are then considered, and its pathology investigated. I have then furnished the materials for its diagnosis, prognosis, and treatment; occasionally illustrating points of special interest by the introduction of cases. In the first chapter I have given, so far as it was necessary for my present purpose, an outline of the existing state of physiology and pathology as applied to the interpretation of diseases of the nervous system. I have occasionally quoted authorities for the sake of enforcing particular points; and a list has been furnished of the editions

of books, papers, &c., to which reference has been made: but it has not been my intention to supply a history of the development of either opinion or practice with regard to the disease in question; as such a proceeding would have occupied much space, which could be more profitably employed by the record of facts which have been observed by myself. An examination of the table of contents will, I think, render still more apparent the general plan and order of the book.

The method employed in almost the whole of this work has been "the numerical," and upon the application of this method to the investigation of disease I wish to make a few remarks. Statistical propositions represent only fractions of the truth which lies beyond them, and the whole of which cannot yet be expressed; they are of value, inasmuch as they tell us what fraction we have obtained. They are of especial use as a means of pointing out and correcting erroneous impressions; of indicating the direction in which true principles or laws may be discovered; and of so registering the facts we have observed, in regard of two or more groups of natural objects, that we may satisfactorily compare and contrast these, and become acquainted with their mutual relations.

If a general principle is enunciated, such as the prevalence of rainy weather after a wet St. Swithin's day, some belief in its correctness may become popular; and numberless individuals,including the most weatherwise of the neighbourhood, and "the oldest inhabitant" of the town,-may so throw their learning and experience into the scale as to turn the balance in the direction of the generally-received belief. The young inquirer, who, from his few years of individual observation, is disposed to bring a verdict of "not proven" against the consequences of St. Swithin's tears, is either laughed to scorn, or "put down" as an insolent sceptic who would dare to question the wisdom of those sages who knew all about St. Swithin before the young would-be philosopher was born. But statistics may show, as they have recently done, that, at all events at Greenwich, a

proposition, the very opposite of that usually entertained, is true. Similar demonstrations have occurred in the history of medicine, and I refer for illustration to that most masterly employment of the numerical method in the investigation by Dr. Walshe, of the hereditary character of phthisis, and of the laws of hæmoptysis.†

In the following work I have shown that some general principles enunciated with regard to epilepsy, not only have exceptions, but that these are more frequent than are the examples of their correctness; and that therefore those so-called "general principles" are demonstrably wrong. Further, the numerical method may be so employed as to show that a principle, antagonistic to that which is generally received, has greater claims to reception; and thus evidence of a positive character may be furnished against the adoption of the received opinion. In this manner statistics have been frequently employed in the course of the present researches upon epilepsy.

It often happens that a startling fact makes so strong an impression upon the mind of its observer, that he attaches to it an importance far greater than it deserves, and this to the depreciation of other and more common occurrences. For example, a certain "cause" has been seen followed by a particular disease; the relation has been carefully established, and the circumstances have been so forcibly impressed upon the mind that the observer is prone to look at other cases in the light of this one; and, as it often happens, when he cannot ascertain any facts in support of his favourite notion, to supply them from imagination or suspicion, and construe all that he does observe into a form which squares with his own theory. We have a natural repugnance to the admission of our own ignorance or error; it is more pleasant to suspect wilful concealment on the

* Nevertheless, it is interesting to observe,-in illustration of another use of the statistical method, -the great prevalence of rainy days at this particular period of the year. This fact, which had been frequently observed, is established by the numerical examination which, at the same time, explodes its popular but erroneous interpretation.

+ British and Foreign Medico-Chirurgical Review, January, 1849.

part of a patient, or even some freak or lusus naturæ, than that our own beautiful generalisation is at fault; and hence, sometimes, we turn away from facts which speak against us, to luxuriate in the society of those which seem all made to our own order. Rejoicing in the thought that we are the "interpreters of Nature," we nevertheless wish that her utterances may be upon our side; and we are disposed not merely to convey, but to parody and convert her teaching. With all reverence to the great men who have built up the science of medicine, it must be admitted that many of their doctrines have had no firmer foundation than that which I have described, and that yet these doctrines have passed current in the minds of their followers for generations; they have been accepted without question, and acted upon without fear; and thus the trustworthy beliefs of the day are compounded with the false notions of tradition as well as with the hasty generalisations of the individual. Now, for the purpose of correcting these, or of showing their true value, when they have any, no method is more convincing than the numerical.

In the employment of statistics with the intention of indicating the direction in which true principles or laws may be discovered, we exhibit perhaps their highest use; but it must be remembered that numerical statements of per-centage and proportion, although they may be perfectly correct so far as they extend, are yet not of the nature of vital or pathological laws.* If it is said that fifty per cent. of the cases of a particular disease, such as epilepsy, present a special symptom, and that fifty per cent. do not, it is not to be regarded as a law of that disease that one half of its examples should differ from the other half. The proper conclusion is, either that the symptom in question, when present, was not essential to the fact of epilepsy, or that the cases from which it was absent were not true examples of that disease. Upon the one supposition, half the cases presented something more than epilepsy; upon the other, half the cases presented less. Either the symptom is

* See Facts and Laws of Life, an Introductory Lecture, by the Author.

a

complication" in fifty per cent., or fifty per cent. of the cases are "abortive," or imperfect. Yet, notwithstanding this constant result of the use of statistics in pathology, it is convenient to retain, as we do, many names-of more or less vague meaning and applicability-to denote diseases which we feel are as yet very inadequately described. It is always highly undesirable to change the names of things, and unless change is imperatively demanded it should be avoided; but the application of the numerical method to the study of disease must result, occasionally, in the disturbance of our nomenclature.

The names by which the greater number of diseases are known express either their essential or prominent symptoms; or some structural condition supposed to be primary to, or causative of the symptoms; or the general sum of all the morbid changes. Under these names, and therefore upon different principles, groups of cases are formed; but each case, in addition to that which it is necessary for it to possess in order to be placed in this or the other category, presents other phenomena than those implied in the name, and than those essential to its position in the group. That which is essential is no matter for statistics in the sense here intended; in every case of pneumonia there must be inflammation of the lung; but in every case of pneumonia there need not be pain in the chest : it is less common than is that symptom in pleuritis; and how much less common it may be highly important for us to know. The diagnosis among many allied diseases is often framed upon a consideration of the proportion borne to each other by symptoms which are, more or less, common to them all; and the diagnostic value of symptoms is in direct relation to the definiteness of our information about them. Such definiteness can only be gained by the numerical method.

The diagnosis of a case is, practically, the process by which, knowing some of its symptoms or phenomena, we infer the others; theoretically, it is the discovery of its primary or essential fact. It is the deductive application of pathology to the particular instance in question; the name we assign to the