3. Wounds of the joints accompanied by grooving of the articular extremities of the bones.

4. Complete perforating tracks through the articular ends of the bones, crossing the joint cavity in various directions.

5. Comminuted fractures of the terminal parts of the diaphyses extending into joints.

Of these several classes, the first was of the least prognostic importance. In the absence of bone injury the wounds usually healed without any obvious ill effect beyond the transient effusion into the joints of a mixture of blood and synovial fluid. When suppuration of the wound in the soft parts occurred, however, the remarks made as to the injuries classed under the third heading also apply here in a lesser degree.

With regard to the retention of the bullet, in the case of bullets of small calibre this was a distinctly rare occurrence. I never happened to see an instance of retention of either a Mauser or Lee-Metford bullet in an articulation. It is only possible with bullets practically spent, or travelling at a very low rate of velocity and making irregular impact.

The influence of both volume and velocity of flight was well illustrated by my own small experience of retained bullets. In one case a Martini-Henry was found impacted between the femoral condyles, having slipped in beneath the margin of the patella. It caused a semiflexed position to be assumed by the joint, and was removed by Mr. Brown in No. 1 General Hospital at Wynberg. The second instance I saw in the Portland Hospital at Bloemfontein in a patient of Mr. Bowlby's. The bullet was a Guedes, a form which has been already described as possessing low velocity and deficient power of penetration; beyond this, in the particular instance irregular impact was evidenced by the presence of a large irregular contused wound of entry over the tuberosity of the tibia.

The presence of the bullet in the knee-joint was later determined by the X-rays, and Mr. Bowlby removed it successfully. Seven months later the range of movement was nearly normal.

I may add that I saw several instances of large leaden bullets lodging in the popliteal space, and a comparatively insignificant number of bullets of small calibre in the same situation. This was very striking, in view of the immense relative frequency of use of the latter forms. There is no doubt, moreover, that small bullets rarely lodge even in the neighbourhood of joints, unless at the distal end of a long track. To take the extreme example of large bullets, those employed as shrapnel, in comparison with the frequency with which wounds were produced by them, bullets lying at the bottom of short tracks in the neighbourhood of joints were not uncommon. Thus I saw one lying over the hip-joint, and another in close proximity to the shoulder capsule.

Wounds of the third class, where the bones had been superficially grooved, were in some respects the most serious. This was especially so in the knee and ankle joints, and some cases will be quoted later under the heading of the special joints to illustrate this point. Danger only arose in the event of suppuration; and here the presence of the long oblique superficial track in a neighbourhood liable to comparatively free movement was the important element. Such tracks usually opened the synovial sac more extensively than direct perforating wounds, and if suppuration occurred in any portion of the track, the pus was very liable to be sucked into the joint on any free movement. The presence of fine splinters of the bone displaced in the production of the groove was also a special character of wounds of this class. Another point worthy of mention is that in these cases it was not always easy to be quite certain whether the joint cavity had been implicated or not, since cases often occurred in which, although the bones had been grooved, the joint cavity escaped. The indication, however, was to consider any wound in the immediate proximity of a joint as perforating until it was healed. This course was the more easy to take, since a large proportion of such wounds were accompanied by some degree of synovial effusion, even when the neighbouring joint had escaped puncture.

Wounds of the fourth class, although the most highly characteristic of the form of accident, were in many instances the most favourable in regard to their course. The tracks might course directly across the joint in any direction, or they might course obliquely, traversing either one or both the component bones. In the latter case the exit might be in the diaphysis, and be accompanied by the separation of an exit fragment such as is illustrated in fig. 52, p. 169. The particularly favourable character of the direct transverse and antero-posterior wounds depended on the slight amount of splintering of the bones, the limited nature of the opening into the joint, and the shortness of the tracks in the soft parts, which ensured that, even if infection did occur, the resulting pus was near the surface, and generally spread in that direction and escaped.