Pierret, is produced by a process of secondary ascending degeneration. While the theory of the secondary degeneration of Goll's columns was tacitly accepted, at this time, by a large number of neurologists, these lesions have been carefully studied of late years, and the accuracy of the original investigations clearly demonstrated.

The idea of a primitive sclerosis of the posterior columns, which made its evolution in situ and of its own will, did not, however, seem doubtful to clinical observers who followed experimental physiology. Waller's experiments were, in fact, a contradiction to the generally accepted idea. Vulpian, in 1879, expressed himself as follows: "We must admit that the alterations of the posterior roots are primitive. It may be that they are so only in a certain measure."

According to this, Vulpian was not far from admitting that in tabes the lesion of the posterior cornu is the consequence of the alteration of the roots. At this time embryology had not shown that the fibres of the posterior cornu are developed at the expense of the spinal ganglia, and the study of the posterior fibres had not yet been begun by section of the roots. Only one thing prevented Vulpian from admitting that the lesion of the cord in tabes was consecutive to that of the roots; that is, that there existed lesions of the cord without alteration of the roots. This question of the rôle played by the posterior roots was left undisturbed for some years, until Leyden, in 1889, took up his former conception, to the effect that this disease is an affection of the sensitive fibres of the cord.

We first promulgated our theory in 1889, in a series of lectures before the Faculty of Medicine. We held that tabetic (ataxic) lesions were only the prolongation of the process into the posterior cornu of the corresponding roots. We announced this theory after a study of the topography of the lesion, and of the proportion, which is constant, in the cases of tabes examined especially in relation to possible alterations of the roots and the posterior cord. We also compared the work of Tooth, who showed by experimental physiology that the lesions of the cord were as we have stated. His, by embryonic studies, has proved that the posterior columns are developed at the expense of the spinal ganglia. So we concluded that the lesions of tabes were the consequence of the lesions of the posterior roots. In 1890 we had occasion to say, in a work on paralysis during ataxia, "To-day tabes appears to be less and less a malady that is localized in the spinal cord. The peripheral nerves, both sensory and motor, are constantly found altered in character; besides, the lesions of the posterior columns are always proportionate to the lesions of the corresponding roots. In

VOL. I. Ser. 4.-8

other words, there is nothing to prove that the spinal lesion is primitive. It is most likely secondary, being a consequence of a neuritis of the posterior roots."

This new conception of the malady has met with some opposition. Dr. Babinski, in his lessons, in Professor Charcot's wards, on peripheral neuritis, said that "sclerosis of the posterior columns in certain cases is preceded by the lesions in the posterior roots." Blocq gave the same reason for opposing our ideas. But Marie and Redlich, after twenty post-mortems, found that we were correct. Neither of these authors had heard of the theory which we had already advanced, but they came to the same conclusions.

In order to show that the lesions of the cord in ataxia are simply those of the posterior roots, we will compare them with those produced by experimental physiology. Taking first the case of compression of the terminal nerves of the cord in men, following a tumor or even a traumatism of the end of the vertebral column, we find that the lesions occupy the posterior columns only, and constantly present the same topography.

Here you notice that Goll's and Burdach's columns are totally sclerosed in the lumbar region, just as in tabes. In the dorsal region the columns of Goll are still in the same state, but those of Burdach are so only in their internal half. In the cervical region Burdach's columns are normal, and Goll's are attached only in their posterior portion. So that in compression of the cauda equina the posterior roots are altered in character, and the sclerosis which occupied the whole of the posterior column in the lumbar region is limited more and more to the posterior half of the columns, until, in the cervical

Root paralysis of brachial plexus: compression Root paralysis of brachial plexus, cervical of first and second pairs, left side (Pfeiffer). region (Pfeiffer).

region, it no longer occupies any part but Goll's columns, and here only the posterior half. Compare this with Figs. 4 and 5, in tabes, and you will see that there is no real difference between them.

If you compare the accompanying drawings (Figs. 6 and 7), taken from Pfeiffer (1891) with the preparations of cervical tabes, you will

see the relation of these cases. Pfeiffer describes a case of radicular paralysis of the brachial plexus, called in German "the Klumpke type."

You will notice (Fig. 6) that the zone of degeneration occupies the external portion of Burdach's columns, just at the level of the posterior roots, and as the preparation is taken from higher up (Fig. 7) in the cervical region, the columns of Burdach are intact in the outer three-quarters, and altered only in the inner quarter, which is close to Goll's columns.

In one word, the topography of this lesion is the same as that of cervical tabes, and experimental physiology shows us that in old cases of this kind where the columns of Goll take part in the lesion the posterior parts of these columns remain clear of trouble.

We know from the experiments of Wagner, Tooth, Singer, Munzer, Barbacci, Oddi, Rossi, and Berdes that the posterior columns of the cord are affected.

Why does this alteration occur in the roots? It must be admitted that we do not know much about it. From this comparative study but one definite conclusion can be drawn,—namely, that in ataxia the spinal lesions are not primitive, but are a systematized sclerosis of the posterior roots, following the intra-medullary tract, which we believe to be primitive in character.

SIMPLE NEUROMATA.

CLINICAL LECTURE DELIVERED AT THE MIDDLESEX HOSPITAL, LONDON.

BY J. BLAND SUTTON, M.D.,
Assistant Surgeon.

GENTLEMEN,-It is important to remember that the term neuroma is often employed clinically in the same sense in which it was used by Odier in the beginning of this century, to indicate a tumor proceeding from a nerve. More accurate knowledge of the minute structure of tumors connected with nerves has taught us that they may be sarcomata, or fatty tumors, or be composed of tissue identical in structure with the endoneurium and perineurium, or consist largely of myxomatous tissue.

Up to the present time, aided by all the best methods of histological research, no one has detected a tumor (apart from amputation bulbs) connected with a nerve composed of new-formed nerve-fibrils. Taking these facts into careful consideration, a neuroma should be defined as a tumor growing from and in structure resembling the sheath of a nerve.

A neuroma may grow from the sheath of any cranial or spinal nerve; usually the nerve is spread over the tumor like a strap; the nerve may traverse the neuroma,—this is rare,—or it may grow within the nerve and spread out its fasciculi like the ribs of an umbrella or a fan. In shape they may be rounded, obovate, or like spindles. The roots of the spinal nerves are sometimes beset with neuromata which take the form of ring-like segments imperfectly demarcated, so that they resemble the annulated rootlets of ipecacuanha.

In the early stages the tissue of a neuroma is dense and resembles the tough tissues of neurilemma: all neuromata are furnished with capsules continuous with and derived from the sheath of the nerve from which they arise.. Later, parts of these tumors undergo degeneration, and in large tumors the central parts liquefy. Thus the terms fibromata, myxomata, and cystic fibromata often applied to these

FIG. 1.

Nerve.

Capsule.

Cavities in the neuroma.

tumors are of no taxonomic value. They may all be classed as simple neuromata or neuro-fibromata. There is a very extraordinary species which is known as plexiform neuromata; but with these we have no concern at present, nor with those rare cases in which neuromata occur in scores on the nerves of the same patient; but to-day I must limit my remarks to the simple species of neuroma of which several examples have presented themselves in my practice, and some which you have had opportunities of examining. Bear in mind, neuromata are not common tumors: if you take our hospital reports you will find that in this institution, where we deal with large numbers of tumors, the average number of nerve-tumors is not more than two annually. Consequently, it is not unusual when a neuroma comes to hand for mistakes to be made in diagnosis: hence I propose to show you how errors arise, and how important it is to be vigilant when removing apparently simple tumors lying in the track of large nervetrunks. The first neuroma I ever removed was of interest from this point of view. A young woman came to the outpatient room for advice concerning a small swelling on the back of her right wrist, situated near the styloid process of the radius; this swelling was smooth, rounded, non-adherent to the skin, painful when pressed, and exhibited all the clinical characters of a synovial cyst (ganglion), except that it did not disappear when the wrist was extended. From a superficial examination I concluded that it was a cyst arising as a diverticulum from the synovial membrane between the radius and the first row of carpal bones, and, as it had existed two years, I thought this would explain the thickness of its walls and its non-disappearance when the wrist was extended. I punctured it with a slender knife, but this had no effect on the swelling, and did not give the patient much pain it was clearly a solid tumor. Cocaine was injected into the skin covering the tumor, which was then exposed by a free incision. It was discovered to be a neuroma connected with the radial nerve as it

A neuroma from Scarpa's space connected with the trunk of the an

terior crural nerve. The tumor has a

distinct capsule; the cavities are due

to degenerative (myxomatous) change.