About eight years ago, whilst examining the ancient British camps on the South Downs, I chanced to discover in the camp of Cissbury, near Worthing, a large flint factory of the neolithic age. There were some sixty or more pits from which flints had been obtained from the chalk, and these pits were full of the débris of the flint-workers. The factory was of the neolithic age, the most characteristic tool of which is the flint celt, a form which differs but slightly from the oval or almond-shaped palaeolithic form, but the cutting edge of which is more decidedly at the broad end. The débris, some six hundred or more specimens of which were collected, consisted chiefly of these celts in various stages of manufacture.
If any one will attempt to make a flint celt, as I have done sometimes (and Mr. Evans, from whom I learnt that art, has done frequently), he will find that it is difficult to command the fracture of the flint with certainty; every now and then a large piece will come off, or a flaw will be discovered which spoils the symmetry of the tool, and it has to be thrown away. In arranging and classifying the remains of this flint factory, I found that all the palaeolithic forms were represented by one or other of these unfinished celts, so much so as to make it doubtful whether some of them may not actually have been used like them.
A celt finished at the thin end, and abandoned before the cutting edge was completed, represented a tongue-shaped palaeolithic implement; a celt finished only on one side represented a palaeolithic side-tool; and a celt rudely chipped out, and abandoned before receiving its finishing strokes, represented almost exactly an oval palaeolithic tool, only differing from it in being somewhat rougher, and showing evidence of unfinish.
Taking a lesson then from this flint-worker’s shop of the later neolithic age, we see how the earlier palaeolithic forms originated. They were not designed outright, as the nineteenth-century man would have designed them for special uses, but arose from a selection of varieties produced accidentally in the process of manufacture. The forms were also suggested by those of the nodules out of which they were made. We see, by examining the outside surfaces that were left on some of them, how a long thin nodule produced a long thin celt, a broad thick nodule a broad thick celt, and so forth. Indeed, so completely does the fabricator appear to have been controlled by the necessities of his art, that in tracing these successive forms one is almost tempted to ask whether the principle of causation lay mostly in the flint or in the flint-worker, so fully do they bear out the statement of Dr. Carpenter and the other physiologists, that nothing originates in the free will of man.
Plate I.
Plate II.
On these two diagrams (Plates I and II) I have shown how, from the same form of palaeolithic implement already described, the more complex forms of the spear and axe-blade of the subsequent periods were developed. The point developed into a spear, and the broad end into an axe-blade. You will see by reference to Plate I that the oval tool of the drift suggested the smaller leaf-shaped spear-head of the early neolithic age. This, by a gradual straightening of the sides, became the lozenge-shaped form, which latter developed into the barbed form, and this last into the triangular form, which consists of barbs without a tang.
On the other hand, this same oval tool of the drift (Plate II), when used as an axe-blade with the broad end, became the celt of the neolithic period, chipped only at first and subsequently polished. This gave rise to the copper celt of the same form having convex surfaces, which grew into the bronze celt with flat sides. Then the bronze celt was furnished with a stop to prevent its being pressed too far into the handle by the blow. Others were furnished with projecting flanges to prevent them from swerving by the blow when hafted on a bent stick. Others had both stops and flanges. By degrees the flanges were bent over the stops and over the handle, and then the central portion above the stops, being no longer required, became thinner, and ultimately disappeared, the flanges closed on each other, and by this means the weapon grew into the socket celt. On this socket celt you will see that there is sometimes a semicircular ornamentation on each side. This semicircular ornament, as I pointed out in a paper on primitive warfare read some time ago, is a vestige of the overlapping flange of the earlier forms out of which it grew, which, like the rings on our brass cannon, are survivals of parts formerly serving for special uses (pp. [182-3] below).