of attachment, but also by the pressure against its inner surface of the head of the humerus around which it is stretched.

The movements which are most frequently concerned in the production of a dislocation are outward rotation and abduction. In the latter the elbow is raised directly outward or outward and forward from the side of the body by the action of the deltoid, the plane in which it moves being more or less exactly that which would be represented by the prolongation of the broad surface of the shoulder-blade. As the movement is made, the head slides downward on the glenoid fossa, the long head of the triceps, the lower part of the subscapularis, and the lower and inner portion of the capsule are made tense, and the movement is arrested when the top of the greater tuberosity comes into contact with the upper margin of the glenoid fossa, and the side of the shaft close below the tuberosity touches the acromion. If the movement is now continued, and the arm raised to the side of the head, it is effected by the rotation of the scapula and the elevation of its outer portion. If, on the other hand, the movement is continued while the scapula is kept stationary, the centre of motion is transferred to the point of contact between the humerus and the edge of the acromion, and the head of the bone is forced downward. against the already tense capsule and ruptures it at its lower and inner portion, there where it presses directly against it.

In outward rotation when the arm is hanging by the side or is but slightly abducted the movement is arrested by the tension of the capsule on the inner side, and at the same time the lower and outer part of the greater tuberosity comes into contact with the outer lip of the glenoid fossa; if the movement is then continued the capsule yields, but the head does not become dislocated unless some other force intervenes to press it inward through the rent that has thus been made.

In all the other movements similar conditions are found, and dislocations following them are less frequent only because the movements are themselves less frequently carried beyond the limits set by the structure of the joint. Thus, adduction and rotation inward are checked by contact of the arm with the body before the capsule is put upon the stretch, and extension of the arm behind the axillary line must be carried very far before a new fulcrum is found, and is also a movement that is rarely produced or exaggerated by external violence.

Statistics. The great frequency of dislocation of the shoulder is fully 'explained by the structure of the joint and by its exposure to the dislocating action of direct and indirect violence. This frequency is so great that dislocations of the shoulder are about as numerous as all the other dislocations of the body combined. The table of statistics given in Chapter I. show percentages varying from 46 to 60 of all dislocations. Malgaigne's statistics of 489 cases contain 321 of the humerus, more than 65 per cent.; Gurlt's collection of 907 cases in the hospitals of Berlin, Paris, and Philadelphia contain 563 of the shoulder, 58 per cent.; Bardenheuer1 saw 20 in a total of 37 cases treated in one year, 54 per cent. Krönlein's statistics, which are especially valuable because they are made up from both hospital and polyclinic records, give a total of 207 dislocations

1 Bardenheuer: Deutsche Chirurgie, Lief. 63 a, p. 279.

of the shoulder, of which 184 were in males and only 23 in females; of Malgaigne's 370 cases 97 were in women; classified according to age and sex they are as follows:

Both show that the injury is rare in youth and old age, is most frequent in middle life, and is much more frequent in men than in women. The relative frequency at the different ages, established by taking into account the percentages of the total population belonging to those ages, differs somewhat from the actual frequency, the maximum being found above the age of fifty years. The proportions calculated from Krönlein's statistics. with the aid of the relative numbers of the population at the different ages, as given in Table V., Chapter I., are 5, 9, 11, and 12 respectively for the decades from 31 to 70. This relatively greater frequency in advanced years is much more marked in women than in men, a fact which is to be explained by the greater exposure to violence incident to the occupations and habits of men in middle life. It indicates, I think, that a much larger proportion of the dislocations in advanced life are due to falls while walking than in middle life, since that is an accident to which both sexes are more equally exposed than they are to others.

The relations pointed out by Krönlein as existing between dislocations of the shoulder and those of the elbow and fractures of the clavicle are interesting. His statistics show that during the first two decades of life, a period in which dislocations of the shoulder are rare, dislocations of the elbow and fractures of the clavicle are most frequent. Thus, of 109 dislocations of the elbow contained in his table, 80 of the patients were under twenty years of age, and of 100 cases of fracture of the clavicle collected by him 70 of the patients were under ten years of age; while of 207 dislocations of the shoulder none of the patients was less than ten, and only 2 less than twenty years old. He thinks fractures of the clavicle are in childhood the equivalent injury of dislocations

of the shoulder by direct violence in middle life, and dislocations of the elbow the equivalent injury of dislocations of the shoulder by indirect violence. He further quotes investigations made by Küstner as showing that separation of the epiphysis in early life, and especially when produced by obstetrical manipulations, is the equivalent of dislocation at other ages, since it is caused by the same mechanism, the forcing of the limb beyond the range of normal motion; the epiphysis separates more easily than the capsule ruptures.

Classification. The head of the humerus in leaving the joint may pass at first upward or downward, backward or forward, and may come to rest in any one of a great number of positions. The classification of the varieties is beset with much difficulty, because of their number, because of the frequency and importance of the secondary displacements, and, last though not least, because of the number of classifications that have already been made and are more or less current. The confusion has been further increased by the application of the same or very similar terms to different varieties by different authors. With the rare dislocations backward, and the still rarer ones upward, there is no difficulty; the uncertainty arises in connection with those in which the head of the humerus has passed across the anterior lip of the glenoid fossa. A brief account of some of the classifications and terms heretofore and still in use will show their differences and resemblances, and may serve as a convenient introduction and preparation for the classification that must follow.

Sir Astley Cooper's classification, upon which those now in use in England and America have been in the main constructed, recognized four kinds of dislocations: 1. Downward and inward into the axilla; 2. Forward, the head of the humerus lying under the clavicle on the sternal side of the coracoid process; 3. Backward; 4. Partial inward, the head resting against the outer side of the coracoid process. It is apparent, from his description, that the first and fourth included the common, frequent cases, those which are now generally termed "subglenoid," or "into the axilla," and "subcoracoid," respectively.

A few years later Malgaigne followed, also with four principal forms, but only one of them the same as Cooper's. His grouping is as follows:

1. Subcoracoid, complete; quite common.

Dislocations into the axilla. 2. Subcoracoid, incomplete; rare.

Dislocations inward.

Dislocations backward.

Dislocations upward

3. Subglenoid; rare.

4. Intracoracoid; most common of all.

5. Subclavicular; rare.

(6. Subacromial; rare.

(7. Subspinous; very rare.

8. Supracoracoid; only two cases known.

All these titles are now in general use; but while the last four, and perhaps the second also, are still used to designate the forms which he designated by them, the others have been used with different, sometimes with widely different, meanings. The first form, the complete subcoracoid, was "characterized by the projection of the head of the humerus in the axilla, and its position exactly below the coracoid process;" it would be included in Cooper's first group, dislocation downward into the axilla.

His second subdivision, incomplete subcoracoid, was the same as Cooper's fourth, partial dislocation inward. His third, subglenoid, was one concerning which he seems to have been far from having very precise notions; he had seen only one case, and had been able to collect only eleven others, and of these the symptoms differed widely, the head of the humerus being described as raising the anterior wall of the axilla in one case, and the posterior in another, as resting against the second intercostal space in one, and against the third in another, and even as having perforated the wall of the chest and lodged within it. The one feature which they had in common, and which he gives as pathognomonic, was that the head of the humerus was not immediately below and in contact with the beak of the coracoid process, but was separated from it by a greater or less interval. Apparently the class was created simply to collect together the odds and ends, the irregular cases that were not subcoracoid; and the idea which suggested the name given to it was that the primary displacement took place more directly downward than in the preceding varieties. It will be seen that the name has since been applied to a very much larger proportion of cases.

His second main division embraced two varieties, the intracoracoid and the subclavicular. Concerning the latter there is no misconception; the term has remained in use, and with the same meaning. The group is made up of those cases in which the head of the humerus has passed entirely to the inner side of the coracoid process, and lies below the clavicle. But the other term, intracoracoid, is generally employed in a much more restricted sense than by Malgaigne. By it he designated the greatest number of dislocations, more than two-thirds of those he saw at the Hôpital St. Louis; he applied it to those in which the head of the humerus, while still remaining under the coracoid process, overlapped it on the inner side by more than half its own diameter. cases are now termed subcoracoid, and only those in which the head has passed almost, if not entirely, to the inner side of the process are called intracoracoid.

Most of such

The tendency of the more recent French and German writers is to make a single group of all the dislocations in which the humerus passes to the anterior side of the scapula, containing four or more subdivisions or varieties, two of which, the subclavicular and intracoracoid, in the narrower sense, are accepted by all. Of the remaining two principal ones, the subcoracoid and the subglenoid, the former is made to include the great majority, and the subglenoid is either closely and distinctly restricted to the very rare cases in which the head of the humerus is displaced directly downward upon the tendon of the long head of the triceps, or Malgaigne's grouping is accepted with all its diversities and vagueness. In the former case the group is removed from the principal division of "anterior" or "præglenoidal" dislocations, and made to form by itself another principal division, termed "dislocations downward."

The English and American writers, as a rule, divide the same cases into subglenoid and subcoracoid, basing the distinction between them upon the clinical feature of the greater or less facility with which the head of the humerus can be felt in the axilla; those in which it is more prominent in the axilla are "subglenoid," those in which it is more prominent

behind the anterior wall of the axilla, close beneath the coracoid process, are "subcoracoid." The objections to this grouping are that it does not sufficiently distinguish between primary and secondary displacements, and that the clinical features upon which it rests present a complete series of intermediate forms, most of which might be as properly placed in one group as in the other. The arbitrariness and uncertainty of the decision are well shown by a comparison of clinical and pathological statistics. Thus, Hamilton and Bryant say that the subglenoid is of more frequent occurrence than the subcoracoid, and Erichsen says that this is the opinion of most English surgeons; while, on the other hand, Flower,' who made an examination of all the specimens contained in the London museums, 41 in number, found that in 32 the dislocation was subcoracoid, and he adds, that of 50 cases recently observed by him in living patients the same was true of "a large majority;" he calls attention to the fact that "the great frequency of subcoracoid dislocation observed in this series [of specimens] does not accord with the descriptions of this injury generally given in the standard surgical works of the country.' A few years later, in the article on Injuries of the Upper Extremity which he prepared in connection with Mr. Hulke for Holmes's System of Surgery, Mr. Flower made a classification in which the influence of this important investigation is apparent. It is as follows:

1. SUBCORACOID. Forward and slightly downward. On to the neck of the scapula, in front of the glenoid fossa, and immediately below the coracoid process. Common.

2. SUBGLENOID. Downward and forward. Head of the humerus in front of the inferior costa [border] of the scapula, below the glenoid fossa. Rare.

His remaining three divisions are Subclavicular, Supracoracoid, and Subspinous, the latter including Malgaigne's sixth and seventh.

Turning now to the pathological data, to the recorded results of postmortem examinations and experiments upon the cadaver, and confining our attention for the moment to the forms mainly in dispute, the dislocations forward (or inward) and downward, and to the points that affect the position of the head of the humerus, the following facts appear:

The head of the humerus, when it passes across the anterior edge of the glenoid cavity, must, as a glance at Fig. 37 shows, move somewhat downward so as to get below the beak of the coracoid process; the position of the limb that most favors the production of dislocation is abduction with or without external rotation. The inner and lower portion of the capsule, being pressed upon by the head of the humerus, tears between the tendon of the subscapularis and the triceps, the rent being small or large and varying greatly in extent and direction in the different cases, but it is always on the anterior and inner side, and the head passes more or less completely through it. If the movement is more directly forward and inward and to a less degree downward, as in dislocations by direct. violence received on the outer side of the shoulder, the head of the bone

1 Flower: Trans. London Path. Soc., 1861, vol. 12, p. 179.

The number is given as 44 in his article on Injuries of the Upper Extremity in Holmes's System of Surgery.