shaft-groove, in the ivory piece which is ingeniously inserted there to form that extremity. This last-mentioned hole is not cylindrical like the one in front, but is so constructed as to allow the shaft-peg to slide off easily. These holes exactly fit two ivory pegs projecting from the harpoon shaft. When the hunter has taken his throwing-stick in his hand he lays his harpoon shaft upon it so that the pegs will fall in the two little holes of the stick. By a sudden jerk of his hand the harpoon is thrown forward and released, the pegs drawing out of the holes in the stick. At the front end of the throwing-stick a narrow piece of ivory is pegged to prevent splitting. As before intimated, this type of throwing-stick is radically different from all others in its adjustment to the pegs on the heavy harpoon. In all other examples in the world the hook or spur is on the stick and not on the weapon.

UNGAVA TYPE.

One specimen from Fort Chimo in this region, southeast of Hudson Bay, kindly lent by Mr. Lucien Turner, is very interesting, having little relation with that from Greenland (which is so near geographically), and connecting itself with all the other types as far as Kadiak, in Alaska (Fig. 3). The outline of the implement is quite elaborate and symmetrical, resembling at the hook end a fiddle-head, and widening continuously by lateral and facial curves to the front, where it is thin and flat. A slight rounded notch for the thumb, and a longer chamfer for three fingers, form the handle. Marks 5 and 6 are wanting. The cavity for the index finger extends quite through the implement, as it does in all cases where it is on the side of the harpoon-shaft groove, and not directly under it. The shaft groove is shallow, and the hook at the lower extremity is formed by a piece of ivory inserted in a parallel groove in the fiddle-head and fastened with pegs. It is as though a saw-cut one-eighth inch wide had been made longitudinally through the fiddle-head and one-half inch beyond, and the space had been filled with a plate of ivory pared down flush with the wood all round, excepting at the projection left to form the hook or spur for the harpoon shaft. This peg or spur fits in a small hole in the butt of the harpoon or spear shaft and serves to keep the weapon in its place until it is launched from the hand. The Ungava spear is heavier than that of the western Eskimo, hence the stick and its spur are proportionately larger. It is well to observe carefully the purport of the spur. A javelin, assegai, or other weapon hurled from the hand is seized in the center of gravity. The Greenland spears have the pegs for the throwing-stick sometimes at the center of gravity, sometimes at the butt end. In all other uses of the throwing-stick the point of support is behind the center of gravity, and if the weapon is not fastened in its groove it cannot be hurled. This fastening is accomplished by the backward leaning of the peg in the Greenland example, and by the spur on the distal end of the throwing-stick in all other cases.

CUMBERLAND GULF TYPE.

The Cumberland Gulf type is the clumsiest throwing-stick in the Museum, and Dr. Franz Boas recognizes it as a faithful sample of those in use throughout Baffin Land (Fig. 4).

In general style it resembles Mr. Turner's specimens from Ungava; but every part is coarser and heavier. It is made of oak, probably obtained from a whaling vessel. Instead of the fiddle-head at the distal end we have a declined and thickened prolongation of the stick without ornament. There is no distinct handle, but provision is made for the thumb by a deep, sloping groove; for the index-finger by a perforation, and for the other three fingers by separate grooves. These give a splendid grip for the hunter, but the extraordinary width of the handle is certainly a disadvantage. There are two longitudinal grooves on the upper face; the principal one is squared to receive the rectangular shaft of the bird spear; the other is chipped out for the tips of the fingers, which do not reach across to the harpoon shaft, owing to the clumsy width of the throwing-stick. In this example, the hook for the end of the bird-spear shaft is the canine tooth of some animal driven into the wood at the distal end of the long-shaft groove.

FURY AND HECLA STRAITS TYPE.

In Parry's Second Voyage (p. 508) is described a throwing-stick of Igloolik, 18 inches long, grooved for the shaft of the bird-spear, and having a spike for the hole of the shaft, and a groove for the thumb and for the fingers. The index-finger hole is not mentioned, but more than probably it existed, since it is nowhere else wanting between Ungava and Cape Romanzoff in Alaska. This form, if properly described by Parry, is between the Ungava and the Cumberland Gulf specimen, having no kinship with the throwing-stick of Greenland. The National Museum should possess an example of throwing-stick from the Fury and Hecla Straits.

ANDERSON RIVER TYPE.

The Anderson River throwing-stick (and we should include the Mackenzie River district) is a very primitive affair in the National Museum, being only a tapering flat stick of hard wood (Fig. 5). Marks 2, 3, 4, 5, and 6 are wanting. The index-finger cavity is large and eccentric and furnishes a firm hold. The shaft-groove is a rambling shallow slit, not over half an inch wide. There is no hook or spur of foreign material inserted for the spear end; but simply an excavation of the hard wood which furnishes an edge to catch a notch in the end of the dart. Only one specimen has been collected from this area for the National Museum; therefore it is unsafe to make it typical, but the form is so unique that it is well to notice that the throwing-stick in Eskimoland has its simplest form in the center and not in the extremities of its whole area. It is as yet unsafe to speculate concerning the origin of this implement. A rude form is as likely to be a degenerate son as to be the relic of a barbaric ancestry. Among the theories of origin respecting the Eskimo, that which claims for them a more southern habitat long ago is of great force. If, following retreating ice, they first struck the frozen ocean at the mouth of Mackenzie's River and then invented the kyak and the throwing-stick, thence we may follow both of these in two directions as they depart from a single source.

POINT BARROW TYPE.

Through the kindness of Mr. John Murdoch, I have examined a number from this locality, all alike, collected in the expedition of Lieutenant Ray, U.S.A. (Fig. 6). They are all of soft wood, and in general outline they resemble a tall amphora, bisected, or with a slice cut out of the middle longitudinally. There is a distinct "razor-strop" handle, while in those previously described the handle is scarcely distinct from the body. Marks 3, 4, 5, and 6 are wanting. The index-finger hole is very large and eccentric, forming the handle of the "amphora." The groove for the harpoon or spear-shaft commences opposite the index-finger cavity as a shallow depression, and deepens gradually to its other extremity, where the hook for the spear-shaft is formed by an ivory peg. This form is structurally almost the same as the Anderson River type, only it is much better finished.

KOTZEBUE SOUND TYPE.

The Kotzebue Sound type is an elongated truncated pyramid, or obelisk, fluted on all sides (Fig. 7). The handle is in the spiral shape so frequent in Eskimo skin-scrapers from Norton Sound and vicinity, and exactly fits the thumb and the last three fingers. Marks 5 and 6 are wanting. The index cavity is a cul de sac, into which the forefinger is to be hooked when the implement is in use. Especial attention is called to this