the density of the bone, and it may be impossible to reach them either by drilling through the alveolar margin or by puncture through the inferior meatus of the nose.

In my clinical experience these small cavities have only been met with in association with a thin narrow face. Macdonald states that the antrum is often small in cases of atrophic rhinitis, and certainly in these cases the outer wall of the inferior meatus of the nose may seem to curve more outwards, but the loss of width thus produced is more than compensated for by increased width of the face.

The interior of the antrum is often more or less divided up by bony septa and membranous folds. Zuckerkandl states that it may be divided into two separate cavities, but this is extremely rare. I have one specimen showing such a condition, and Logan Turner has once met with an almost complete septum. Grüber has found five complete divisions in two hundred specimens. Cryer regards these second cavities as ethmoidal cells. Frequently the septa are so large that their removal is required to obtain free drainage in cases of suppuration. The antral cavity may also be irregularly enlarged by a local bulging out of its walls in almost any direction. The ridge of bone carrying the infraorbital vessels and nerve almost constantly cuts off a part of the antrum in its upper and inner part, thus forming the infraorbital cell (Cryer).. Development of the Antrum. The antrum is present as a very small slit-like cavity at birth, being 1-3 mm. in height by 7 mm. in anteroposterior diameter. It gradually increases in size as the eruption of the milk teeth proceeds. At six years it forms a more or less round cavity lying opposite the middle meatus of the nose, measuring 5-8 mm. in diameter, and extending outwards beneath the orbit as far as the infraIt is separated from the hard palate and alveolar process by a mass of cancellous bone in which lie the germs of the permanent teeth (see Fig. 9). This general arrangement continues until the eruption of the permanent teeth, that is until about the seventh year, when the cavity commences to extend downwards towards the hard palate and alveolar process, gradually attaining its pyramidal shape and adult dimensions with the complete development of the upper jaw.

The structure of the upper jaw at the seventh year, consisting mainly of a large mass of cancellous bone, is well shown in Fig. 9. The development of the antrum has an important surgical significance, especially in reference to certain conditions which have been described as empyemata of the maxillary antrum occurring at early ages. These are really cases of osteomyelitis of the upper jaw, the true pathology of which has been completely misunderstood (see Chap. XVIII). It is also necessary to bear in mind the development of the upper jaw in order to understand the irregularities in its formation produced by obstruction to nasal respiration (see Chap. IV)..

Quoted by Turner.

Frontal Sinus. The frontal sinus is formed by the separation of the inner and outer tables of the frontal bone in the supra-orbital region, and consists of a vertical portion extending upwards on to the forehead, and a horizontal portion extending backwards over the orbit. It is an irregular, pyramidal cavity, with its apex at its highest point in the frontal bone, its base formed by the roof of the orbit, and with three walls, an anterior, a posterior, and an internal. The last-named wall separates the two frontal sinuses, and is consequently often termed the sinus septum (see Fig. 10).

The anterior wall is convex outwards, both from above downwards and from side to side. Its thickness varies considerably in different skulls, ranging from 2-6 mm. It is often stated, and the statement is emphasized, that this wall contains no diploë, or that traces at most are present; but repeated observations have convinced me that diploic tissue is constantly present, although often in small amount (see Fig. 14).

The posterior wall is usually very thin, -1 mm. in thickness, and is composed of dense brittle bone with no cancelli. It is convex towards the sinus. Its posterior surface is indented by the convolutions of the frontal lobe of the brain.

The internal wall, or sinus septum, is very thin, but almost always complete. In one hundred and four skulls Tilley found it complete in all. Lothrop saw it perforated only twice in one hundred and eighty specimens. Turner found it always complete, and Bosworth is evidently in error when he states that the sinuses frequently communicate. In cases of suppuration,

however, the septum may become perforated, and this is probably not very rare. I have three times seen this; and Luc, Tilley, and others have recorded a similar condition. The lower part of the septum is usually in the middle line as it extends upwards it may be very deflected, or become almost horizontal. If in a skull such as that shown in Fig. 12, an attempt had been made to perforate the right frontal sinus above the supraorbital margin, although the opening was made well to the right of the median line, it would have been quite possible to have opened the left frontal sinus. In rarer cases the septum is twisted so that one cavity may partially overlap the other.

The inferior wall consists of a large outer or orbital portion forming the roof of the orbit, and a small inner or nasal portion forming the roofs of some of the anterior ethmoidal cells. The orbital portion is very thin, is convex towards the sinus, and internally bends abruptly downwards to join the lamina papyracea of the ethmoid. It is perforated by a small vein. near its inner end. The nasal portion of the floor is a small irregular surface composed of very thin bone separating the sinus from a variable number of anterior ethmoidal cells. The inferior wall of the frontal sinus is the thinnest, and is consequently the commonest seat of perforation or bulging, which often occurs in the neighbourhood of the small vein above mentioned. It is here also that tenderness on pressure should be sought

for. It is probable that the nasal portion of the floor is often perforated in disease of the sinus.

Size of Cavity. The frontal sinuses vary much in size. They may be very large and extend outwards as far as the external angle of the orbit, upwards for one, two, or even two and a half inches, and backwards to the posterior part of the roof of the orbit. Turner gives as an average measurement: Height 31.6 mm., breadth 25.8 mm., and depth, measured along the roof of the orbit, 18 mm. The depth behind the supra-orbital margin is much.

less, averaging 4-8 mm. The frontal sinuses may be very small; they may be limited to small cavities situated at the internal angle of the orbit, the vertical portions being missing: occasionally both sinuses are entirely absent. The sinuses are frequently asymmetrical, a large cavity on one side being associated with a small one on the other: or only one cavity of a variable size may be found. A single large sinus. may extend from one supra-orbital region to the other; the true nature of the cavity will be indicated by the presence of a single ostium communicating with only one nostril. Lothrop in two hundred and fifty specimens found both sinuses invariably present in the orbital region,

but in 3 per cent. the vertical portion was absent. He, however, describes one specimen with a large sinus on one side and apparently none on the other.1 Logan Turner, in two hundred and forty European skulls, found one or both sinuses absent in forty-one; that is in 17 per cent. Of these, in eighteen, that is in 7.5 per cent., both sinuses were absent, and in twenty-three, or 9.5 per cent., one sinus was absent. Tilley 2 found entire absence of the sinuses twice in 120 skulls, in one case in association with prominent brows. Kicer found both sinuses absent five times, and one sinus absent seven times in 195 skulls. Max Scheier 4 found no frontal sinus twice in 100 skulls. On the whole, the sinuses seem to be larger in males than in females, and are smaller and more often absent in races with receding foreheads; but the size of the cavity bears no constant relation to the prominence of the supra-orbital area.

3

In the living body two means may be adopted to ascertain the existence of sinuses-transillumination and the Röntgen rays. Of these the latter is the more reliable: when the sinuses are large their presence can be definitely determined, both by the screen and by photography, but if the sinuses are small, these means may fail. The results of transillumination are very uncertain. Unless a definite cavity can be demonstrated by means of the Röntgen rays, external operation alone can decide its presence or absence.

As in the antrum, bony and membranous ridges partially dividing the cavity are often found: they usually run in a vertical and antero-posterior direction. Very rarely there is a complete septum dividing a sinus into an inner and an outer part, but usually although it may appear complete, there is an opening at the lower part, so that both divisions of the sinus open into the nose through one ostium. Another condition simulating a double frontal sinus is the presence of a large ethmoidal cell extending well outwards over the orbit. One or two cells are normally present near the nasal floor of the sinus, and in rare specimens a cell as large as the frontal sinus itself may be found, and its true nature may be easily overlooked. Instances of true double frontal sinus are recorded by Scheier, Hansen, Pluder, etc., and I possess one specimen showing it on both sides.

The frontal sinuses may have various diverticula or recesses, especially near the outer angle, or there may be a prolongation from the inner end of the sinus into the crista Galli. According to Cryer, the latter is found in 10 per cent. of specimens. These recesses are of considerable surgical importance, as they may be easily overlooked during operations: suppurating foci may be left behind and the operation may prove a failure.

1Annals of Surg. 1898, xxviii. p. 622.

2 Lancet, 1896, ii. p. 867.

3 Laryngoscope, 1899, vi. p. 83.

Archiv für Laryngol., 1902, xii. p. 296, and Archiv Internat. de Laryngol., 1901, xiv. p. 321.

Development. The frontal sinuses are absent at birth, and up till about the sixth year. They apparently begin to develop as a protrusion from the ethmoidal region at about the sixth or seventh year, and slowly spreading between the tables of the frontal bone, attain their full size and dimensions at or about the age of puberty. From nine to twelve years old fair-sized cavities have been seen (Symington).1

Relations. The most important of the external relations of the frontal sinuses are as follows: The outer part of the inferior wall forms the roof of the orbit, and at its inner end is a small depression in which is fixed the pulley of the superior oblique muscle. The attachment of this pulley is often necessarily displaced in operations on the sinus. The anterior margin is crossed at its inner end by the supra-trochlear vessels and nerve, which are necessarily divided in operations, but are of no great importance.

FIG. 13.-A DISSECTION SHOWING THE EXTERNAL RELATIONS OF THE FRONTAL SINUS. The right frontal sinus and infundibulum have been opened from the front. On the left side a skin flap has been turned back to show the position of the pulley of the superior oblique. a. Lachrymal duct; b. commencement of the infundibulum; c. nasal septum; d. pulley of the superior oblique; e. cavity of the frontal sinus; f line of incision in the supra-orbital margin; g. tendon of the superior oblique.

If the sinus extends far outwards the supra-orbital vessels and nerve also come into relation with its anterior wall (see Fig. 13).

The posterior wall is in relation with the dura mater and frontal lobe of the brain for the greater part of its extent, and with the olfactory lobe at its inner border.

The ostium of the frontal sinus is in the posterior part of the nasal portion of the floor, near the septum; being placed at the most dependent part of the cavity, it is well situated for drainage. It is round or oval, averaging 3 mm. in diameter, but varying from 2-8 mm. It leads into the infundibulum, a long curved canal (15 nim.) running downwards and backwards to the anterior end of the hiatus semilunaris This canal is convex forwards and downwards; it is bounded below by the uncinate process and the bulla ethmoidalis, and above by the anterior ethmoidal cells and the lachrymal bone. The canal may be wide and nearly straight, or