Table II.—Analyses of Ancient European Bronzes

This argument, however, is defective when taken to determine the question of the origin of bronze in favour of independent discovery, for we have already seen, in speaking of the stone age,—and I have endeavoured to show that it is a peculiarity observable in the works of all savage and barbarous races,—that being devoid of rule or measure, and having very imperfect means of securing adherence to a uniform standard, their productions are characterized by incessant variations, even in cases where the first idea is known to have been derived from a common source. The variations here shown to exist in the composition of bronze are no greater than are capable of being accounted for by the universal prevalence of a law of variation, resulting from many causes, and amongst others from want of precision, and carelessness, which is a defect common alike to all tyros in their art, whether ancient or modern. It is a fault we have many of us to complain of almost daily in our cooks. A batter pudding is composed of milk, flour, and eggs, in proper proportions, but a careless cook will constantly vary her proportions, and will fail in adjusting her quantities to the total amount; but we must not, on that account, assume that each cook has invented the art of making batter puddings independently. So, in like manner, it is quite consistent with the facts observed even in America, to suppose that the first knowledge of bronze, and of those many features in the civilization of the Mexicans and Peruvians which present such striking analogies to the civilization of Egypt, may have been originally communicated by some casual wanderer or some shipwrecked castaway from the then centres of Eastern culture (for the theory of geographical changes is, of course, out of the question when speaking of the origin of bronze), and that they have varied in their development on American soil no more than might naturally be expected from their introduction to an entirely new and partially civilized race. Such an assumption, though difficult to account for, and wanting in evidence, is more in accordance with the well-known traditions of the Mexicans and Peruvians, who attribute their civilization to the advent of a god; or with that of the natives of Nootka Sound, on the north-west, who state that an old man entered the bay, in a copper canoe, with paddles of copper, and that the Nootkans by that means acquired a knowledge of that metal.

As illustrations of the modern metal-work of the natives of Nootka Sound and its neighbourhood, several examples are given in Plate XIX, figs. 7 to 11. Figures 7 and 8 represent two sides of an iron dagger in the Museum of the Royal United Service Institution. The ornamentation on the handle is that of the natives of the country, but the workmanship of the blade, which is ribbed on one side, appears to indicate foreign manufacture. Figures 9 and 10 are two sides of a copper dagger of the same form; this specimen is now in the Belfast Museum, in which it was deposited in the year 1843 by Mr. A. Thompson, who brought it from the north-west coast of America, and described it as having been fabricated by the Flathead Indians; it is undoubtedly of native workmanship; in both these weapons one side of the blade and handle is concave, the other convex, a form which appears to denote that it was originally taken from some similar weapon of bone or cane. The nearest approach to the form of this weapon in bone, that I am aware of, is that of the Indian ‘kandjar’, a figure of which was given in my first lecture on Primitive Warfare, Plate X, fig. 63. This weapon has also one concave and one convex side, derived from the natural curvature of the bone out of which it is made.

But putting aside American civilization, which, it must be admitted, does in the existing state of our knowledge present great difficulties in the way of those who advocate the theory of a common origin for bronze, and turning our attention to the eastern hemisphere, we find the evidence on this point more satisfactory. We may observe, in the first place, that the area over which bronze has been used for implements appears, in so far as we have at present been able to trace it, to be continuous, extending over the greater part of Europe, Egypt, Assyria, and some parts of Siberia, India, and China, from which latter country some few bronze weapons have lately been added to the British Museum. Mr. Theobald, of the Geological Survey of India, also mentions in a paper read to the Bengal Asiatic Society,[207] that bronze axes are found in the valley of the Irrawaddy, where they are held in such veneration as rarely to be procurable; and Sir Walter Elliot has shown me some bronze implements which he found deep beneath the soil in cutting a canal in the valley of the Ganges. Bronze is wanting in Africa; in America, with the exception of Peru and Mexico; in the north of Sweden and Norway, and, I believe, in the greater part of the northern districts of Russia and Siberia, though with regard to Russian and Siberian bronzes, our information is still very deficient. And here I may observe that I speak only of bronze as applied to tools and weapons; its use for other purposes may have been introduced at any subsequent period of the world’s history; but the presence of a bronze weapon implies either total ignorance, or at least an imperfect knowledge of the means of hardening the more useful metal for this purpose, iron.

Those who wish for more detailed information as to the evidence upon which the succession of the stone, bronze, and iron ages has been determined, would do well to refer to Sir John Lubbock’s remarks upon this subject in Prehistoric Times. It may, however, be useful to enumerate briefly some of the chief points which have been adduced in support of the opinion that the employment of these materials corresponds to successive stages in the development of civilization in Europe. (1) Not only do the Roman writers mention iron as being the metal used by them in their time, but they also speak of its employment by the barbarian nations of the north, with whom they came in contact, and the word ‘ferrum’, iron, was with the Romans synonymous with sword. (2) Although numerous finds of iron implements of the Roman period have been discovered in various parts of the world, there has been no authentic and undoubted instance of a weapon of bronze having been found associated with them, or with Roman pottery or coins. (3) Bronze implements are most abundant in Denmark and Ireland, countries which were never invaded by Roman armies, whilst they are exceedingly rare in Italy. (4) The ornamentation of the bronze implements is not Roman, but pre-Roman in character. (5) On the other hand, the numerous finds of bronze weapons which have been discovered have never been associated with iron, except in cases where the nature of the iron implements shows them to have belonged to a period of transition. (6) The pottery associated with bronze-finds is superior to that found with stone implements, but inferior to that of the iron age, and the potter’s wheel was unknown during the stone and bronze ages. (7) Silver is found associated with iron, but rarely if ever with stone or bronze. (8) No coins or inscriptions of any kind have been found with bronze implements. (9) In the Swiss lakes, settlements associated with stone and bronze have been found near each other, as for instance Moosseedorf and Nidau, 15 miles apart; in the former, bronze is entirely absent; in the latter, it was used not only for articles of luxury, such as might denote a more wealthy class, but also for implements of common use, such as fish-hooks, pins, &c.; it is improbable that so marked a contrast in the civilization of two settlements so close to each other should have existed during the same period. (10) The implements and ornaments of the bronze-finds are more varied in form, showing an advance in art upon those appertaining to the stone age. (11) The bronze-finds are marked by an increase in the number of domesticated animals, and an entire absence of some of the wild animals of the earlier period, and they are also more clearly associated with traces of agriculture. (12) In the Danish peat bogs, successive strata are found overlying each other, denoting changes in the vegetation of the country; in the lowest and earliest are found the remains of pine trees, which now are foreign to the soil; above which are strata in which oak was the prevailing tree, and at the present time the oaks have been superseded by beeches. These successive strata correspond in a general way to successive stages in the civilization of the inhabitants; in the pine-bearing strata, implements of stone are found; with the oak trees, implements of bronze, and higher up, implements of iron. It has also been attempted to trace a somewhat similar succession of periods in the gravels and alluvium of the torrent of Tinière, in Switzerland; but the evidence in this case is not considered so satisfactory as in that of the Danish peat bogs.

In Chaldea, the transition from stone to bronze has been traced by the relics found in the soil; iron being then used only in small quantities, and chiefly for ornaments, as amongst the ancient Britons in the time of Caesar.[208] In Egypt, where both bronze and iron weapons have been found in the tombs, the transition from bronze to iron is marked by the colour of the weapons in the paintings, and dates, according to Sir Gardner Wilkinson, about B.C. 1400. Hesiod speaks of an age of copper, when the ‘black iron did not exist’. Homer also alludes frequently to copper or bronze implements, and when iron is mentioned always speaks of it as requiring much time and labour to fabricate it. Then we have the well-known passage from Lucretius, so often quoted in reference to this subject, in which the three ages of stone, bronze, and iron are mentioned;[209] and Strabo mentions the Lusitanians as being armed partly with copper or bronze weapons.[210]

Many other quotations might be given from ancient authors to prove that the existence of a bronze age preceding the use of iron was known to the ancients, but I will not occupy your time further with this part of the subject, seeing that others far more competent to deal with it than myself have failed to derive much information of value from this source. There is often considerable difficulty in determining the exact meaning of the writers, when speaking of the material of which weapons are composed, the same word being sometimes used to express copper, bronze, and iron. In fact it may, I think, safely be said that, notwithstanding the large amount of useful information that may be obtained from the study of the early writers, there is no more fruitful source of error than the attempt to apply ancient history and tradition to the elucidation of prehistoric events. Modern science, and our fuller appreciation of the value of evidence, have thrown far more light on prehistoric times than ever fell to the lot of the ancients; and it is for us, therefore, to correct their errors, and not to be misled by them.

Professor Max Müller, in the second series of his Science of Language, has, however, drawn some important conclusions on this subject, from the etymology of words representing metal, of which it may be useful here to give a brief abstract. Quoting Mr. E. B. Tylor’s work on the Anahuac (p. 140), he says: ‘The Mexicans called their own copper or bronze tepuztli, which is said to have meant originally hatchet; the same word is now used for iron, with which the Mexicans first became acquainted through their intercourse with the Spaniards. Tepuztli then became a general name for metal, and when copper had to be distinguished from iron, the former was called red tepuztli, and the latter black tepuztli. The conclusion,’ he says, ‘which we may draw from this, viz. that Mexican was spoken before the introduction of iron into Mexico, is one of no great value, because we know it from other sources’; but applying the same line of reasoning to Greek, he says, ‘here, too, chalkós, which at first meant copper, came afterwards to mean metal in general, and chalkeús, originally a copper-smith, occurs in the Odyssey (ix. 391) in the sense of a blacksmith, or worker of iron.’ What does this prove? It proves that Greek was spoken before the introduction of iron. The name for copper is shared in common by Latin and the Teutonic languages, æs, Latin; aiz, Gothic; êr, old high German; erz, modern German; âr, Anglo-Saxon; and the same word is represented in our English word ore. But the words specifically used for iron differ in each of the principal branches of the Aryan family. At the same time the words originally representing copper come to be used for metal in general, and in some cases for iron. In Sanskrit, ayas, which is the same word as æs, came to be used for iron, a distinction being made between dark ayas or iron, and bright ayas or copper. Æs in Latin, and aiz in Gothic, came to be used for metal in general, but was never used for iron. Aiz, however, according to Grimm, gave rise to the Gothic word eisarn, meaning iron. In old high German eisarn is changed into îsarn, later to îsan, and lastly to the modern eisen, while the Anglo-Saxon îsern is converted into îren, and ultimately to iron. The learned Professor sums up his researches on this subject as follows:—‘We may conclude,’ he says, ‘that Sanskrit, Greek, Latin, and German were spoken before the discovery of iron, that each nation became acquainted with that most useful of all metals after the Aryan family was broken up, and that each of the Aryan languages coined its name for iron from its own resources, and marked it by its own national stamp, while it brought the names for gold, silver, and copper from the common treasury of their ancestral home’.[211]

These remarks point to a very remote period, and to an Aryan origin for the first knowledge of copper and bronze, but on the other hand much has been written in favour of a Semitic origin, especially by Professor Nilsson, who believes that he has discovered traces of that people even on the coast of Norway.[212]

The employment of war chariots, which are known to have been used by the Britons, and vestiges of which have been found in their graves, implies, it is said, Semitic influence. Much stress is also laid upon the resemblance of some of the ornaments found on the Danish and other bronzes to those in use by the Phoenicians; more especially the spiral ornaments, which Professor Nilsson traces to that source through the engravings on weapons in the bronze age tumuli. Against this, however, it may be urged that the spiral ornament has a very wide distribution, extending over modern Africa, ancient Egypt, Greece, China, New Guinea, Mexico, and South America, and even to New Zealand and the Asiatic Isles. In illustration of this I have arranged upon Plate XIX a series of illustrations of spiral ornament from various countries, showing how universally it is distributed over the globe. Fig. 12 is from a New Zealand canoe in my collection; Fig. 13, from a club brought from New Guinea by the commander of the ‘Rattlesnake’, in 1849, and now in my collection; Fig. 14, from China; Fig. 15, from ancient Egypt; Fig. 16, from Greece; Fig. 17, from a Danish bronze sword; Fig. 18, from an Irish bronze brooch in my collection; Fig. 19, from the Swiss lakes, figured in Dr. Keller’s work; Fig. 20, an iron ornament in my collection from Central Africa; Fig. 21, an iron ornament on a club, from the Bight of Benin, West Africa, in the Christy Collection; Fig. 22, an ornament on a wooden arrow-head, in the Christy Collection, probably from one of the Melanesian isles; Fig. 23, from Hallstatt; Fig. 24, a cane arrow-head from the Amazons, South America; Fig. 25, a spindle-whirl from Mexico; Fig. 26, on a bronze shield from the Caucasus; Fig. 27, an ornament on a bracelet from Hindustan, in the British Museum; Fig. 28, an ornament carved upon the stones of New Grange, in Ireland; Fig. 29, from a New Zealand canoe. Compare the two last figures with Fig. 30, a stone weight in my collection, lately fished up on the coast of Kent, whilst dredging for whelks; the ornamentation so closely resembles the New Zealand pattern, and at the same time that of the stone carvings of the European tumuli, that considering the circumstance of its discovery, it is purely a matter for conjecture whether it is to be referred to the antiquities of this country, or has been dropped overboard by some vessel returning from our South Pacific colonies. We see from these examples that the spiral ornament cannot be regarded as belonging exclusively to any one race; it is a contrivance derived simply from the coil of string, the source from which, and also from straw plaiting, nearly all barbaric ornamentation had its origin; it is a proof merely of barbaric origin, an evidence of continuity from the earliest periods of art.

Mr. Franks in his remarks at the Paris Meeting of the International Congress of Prehistoric Archaeology, has summarily disposed of the question of Phoenician ornamentation, by observing that the Phoenicians were copyists, taking their style from Egypt, Greece, or Rome, according to the fashion of the period, and that in point of fact a Phoenician style of art has never existed (Compte Rendu, II^me Session, Paris, 1868, p. 251).

Amongst those who have upheld the theory of the origin of bronze from Phoenician sources, may be mentioned Mr. Howorth, in a paper lately published in the Transactions of the Ethnological Society (1868, N.S., vol. vi. pp. 73-100); and Sir John Lubbock, though not committing himself to the same view as regards the origin of bronze, has nevertheless been at the pains of ably defending the ancient authors who speak of Phoenician intercourse with Britain from the attacks made upon them by Sir George Cornewall Lewis (Prehistoric Times, 1869, pp. 59-69).

This being the existing state of our knowledge in regard to the introduction of bronze, and the variety of opinion on the subject being, as we have seen, considerable, the task before us will be to ascertain as far as may be possible, from the implements themselves, the history of their origin, by examining carefully their construction in the various regions in which they occur, and by tracing the geographical distribution of those details of form which show evidence of connexion; thereby to determine, if possible, the sources from which they were derived. Whatever degree of veracity we may be disposed to attribute to early history, we must at least admit that the implements have this advantage over written testimony of any kind, that they cannot intentionally mislead us. If we draw wrong inferences from them, the fault is our own. We shall find the evidence very fragmentary as yet, but sufficient to prove that it affords a valuable source of information whenever sufficient materials are collected to enable us to work out the problem to its legitimate ends.

On the present occasion I propose to confine my remarks to showing, by means of the accompanying table (Plate XVIII), the distribution of some of the commoner varieties of the copper and bronze celt, an instrument which, like its prototype in stone, appears to have been employed both as tool and as weapon for all the various purposes to which it was capable of being turned, and to have been used not merely as a hatchet and battle-axe, but also to have been sometimes hafted on the end of a straight handle, to be used as a spud or crowbar, and even perhaps, as some of the forms appear to indicate, as a spade in tilling the ground.

The table is arranged upon the same plan as Plate XIII of my last lecture, and is intended to serve as a continuation of Plate XII of the same lecture, showing a further development of the same weapon. The successive developments are arranged, in order, by classes from left to right; the several localities are separated by horizontal dotted lines, by means of which are seen the various types prevalent in each locality, in so far as I have been able to obtain drawings from published sources; there can be no doubt, however, that the table is still very imperfect, and that considerable additions may be made to it hereafter. On the left, in Class A, will be found celts with convex surfaces, identical in form to those constructed of stone, the relative antiquity of which is shown by their being almost invariably of pure or nearly pure copper. It has been suggested that this form may have been adopted on account of its being more easily produced by beating the copper, and that its resemblance to the stone celts is not necessarily a proof of age; but there is no reason why Class B should not be as easily formed as Class A by this means, and many are so formed, as may be seen in the table. Moreover, Fig. 3 a is a bronze celt of the earlier form, taken from Prehistoric Times, and as this must have been cast in a mould, its peculiar shape can only be accounted for by supposing it to have been constructed in imitation of the stone celts. In passing from Class B, a gradual development of form may be traced, commencing with a slight stop or ridge across, and rudimentary flanges along the side of the shaft of the blade, developing in size and improving in form, no doubt, as the art of casting bronze became gradually perfected.[213] These stops and flanges are at first raised on the surface of the blade, but by degrees the same purpose is effected by sinking a groove in the blade to receive the handle, thereby economizing the metal, and producing a more symmetrical form; the flanges were at the same time bent over, and ultimately cast with a cavity on each side to receive the handle, and obviate the necessity for binding on the celt with thongs. This led by degrees to the ultimate perfection of the weapon, by the introduction of the socket type, which is associated with weapons of iron, and is sometimes itself constructed of that metal.

The order of development here adopted is in the main that followed by Sir William Wilde, in his Catalogue of the Museum of the Royal Irish Academy, but I have omitted all mention of branch varieties, as they do not serve my purpose of illustrating the continuity of development, though they are valuable in showing the connexion between localities.

Although the course of development appears to have followed the order here indicated, it is not unlikely the earlier forms may have continued in use, and may even have continued to be constructed at the same time as the later forms. The earlier and less complicated types, being easier of construction, and being equally serviceable for some purposes, would continue to be made, in the same way that smooth-bores and rifle-barrels, row-boats, sailing-vessels, and steam-packets, continue to be used simultaneously in our own time.

The progress of development of this weapon will be better understood by a detailed reference to the figures.

Reference to the Figures in Plate XVIII.[214]
COPPER, BRONZE, AND IRON CELTS.

Class A.—Copper celts from various localities, having convex surfaces, in form resembling those of stone.—Figs. 1, 2, and 3, from Ireland, in my collection.—Fig. 3 a, a bronze celt of the same form, from Le Puy, France, Prehistoric Times, p. 27.—Fig. 4, copper celt found at Blengow, Mecklenberg-Schwerin Museum; Horae Ferales.—Fig. 5, copper celt from the lake dwellings of Sipplingen, Switzerland, found embedded in a coating of clay (a mould?). See Keller, The Lake Dwellings of Switzerland, (transl. J. E. Lee, 1866), p. 121, Plate xxix.—Fig. 6, copper celt found in an Etruscan tomb, and now in the Berlin Museum. See Catalogue of the Royal Irish Academy, ‘Bronze,’ pp. 367, 395.

Class B.—Copper and bronze celts from various localities, having flat concave sides, and a rectangular cross section, showing a gradual enlargement of the cutting edge.—Figs. 7 to 12, copper celts from Ireland, in my collection, showing a gradual enlargement of the cutting edge.—Figs. 13, 14, 15, ditto, ditto, of bronze, the sides more concave, and the cutting edge more expanded.—Fig. 16, bronze celt, of similar form, from Denmark (Madsen, Afbildninger af Danske Oldsager og Mindesmærker, Copenhagen, 1872, Heft iii, Fig. 1).—Fig. 17, copper celt from Steinfurt, in the collection of Professor Dieffenbach, at Friedberg, Lindenschmit, Die Alterthümer unserer heidnischen Vorzeit (Mainz, 1864 ff.), Plate 3.—Fig. 18, ditto of copper, found near Mainz, Museum of Mainz, Lindenschmit, Plate 3.—Fig. 19, the same form of bronze, from near Mainz, Lindenschmit.—Fig. 20, the same form of bronze from Italy, British Museum.[215]—Figs. 21, 22, 23, the same form of copper from Hungary, Keller, p. 219, Plate lxviii.—Figs. 24, 25, 26, similar forms of bronze, with rectangular holes, from the Island of Thermia, Greek Archipelago, British Museum.

Class C.—Bronze celts of the same outline as Class B, but having a cross ridge or stop on both faces, to prevent the blade from burying itself in the handle.—Figs. 27, 28, bronze celts from Ireland, in my collection; this form is common to the British Isles.

Class D.[216]—Bronze celts, having four longitudinal ridges or flanges, one on each edge, but no cross stop. The flanges are for the purpose of fixing the blade in a bent handle; they exhibit a gradual development of the flange, and an expansion of the cutting edge, which latter takes a semicircular, and in some cases nearly a circular form.—Figs. 29, 30, from Ireland, in my collection, showing front view and section.—Fig. 31, from Versailles, in my collection, with section.—Fig. 32, from France; with side view; see Matériaux pour l’Histoire de l’Homme.—Fig. 33, from Loyette, Department of Isère, from Horae Ferales, front view.—Fig. 34, from the South of France, British Museum, the blade very circular.—Fig. 35, from Alps [Aps?], in Ardèche, British Museum, the circular form of the blade still more developed. This form appears peculiar to the neighbourhood of the Rhone, Horae Ferales.—Fig. 36, from France; with side view; Matériaux.—Fig. 37, from Denmark, British Museum, of copper; this form is rarely found in copper; with section.—Fig. 38, from Denmark, of bronze, from Madsen, Heft iii.—Fig. 39, from Denmark, with semicircular blade, Madsen, Heft iii.—Fig. 40, from Hessen, now in the collection at Hanover, Lindenschmit, Heft i, Taf. iii.—Fig. 41, from near Baltringen, Lindenschmit.—Fig. 42, from Neinheiligen, in Thuringia, British Museum; with section.—Fig. 43, from the Terramara Beds, Castione, Switzerland; with section; Keller, Plate lix.—Fig. 44, from Unter Uhldingen; with section; Keller, Plate xxix.—Fig. 45, from the Terramara Beds, Castione; with section; Keller, Plate lix.—Fig. 46, from the Terramara Beds, Castione; with section; Keller, Plate lix.—Fig. 47, from Hallstatt, in Austria, von Sacken, Das Grabfeld von Hallstatt in Oberösterreich und dessen Alterthümer (Vienna, 1868), Taf. vii; with side view.—Fig. 48, ditto, ditto, found with the body of a child.—Fig. 49, ditto, the shaft of bronze, and the blade of iron, from Hallstatt.—Fig. 50, the same form in iron, also from Hallstatt, in Mr. John Evans’ collection.—Figs. 51 and 52, similar forms, in bronze, from Italy, British Museum.—Fig. 53, the same form, from Telsch, Vilna, Russia, British Museum; with two sections.

Class E.—Bronze celts having both the cross stop and the longitudinal flanges. In the earliest form, the cross stop and flanges are raised upon the faces of the blade, as in Class D. In the more improved form, the upper part of the shaft of the blade is hollowed so as to answer the same purpose and economize the metal. Figs. 54-8, from Ireland; Fig. 54, with rudimentary stop and flanges, in my collection. Figs. 55 and 56, ditto, with rudimentary stop, the flanges more developed; in my collection. Fig. 57, showing a development of both stop and flange, ditto, ditto. Fig. 58, showing the stop and flange further developed, and the metal of the upper part of the blade slightly sunk, ditto, ditto. Fig. 59, a further development of the same, the metal of the upper part of the shaft of the blade reduced to a minimum.—Fig. 60, the same form as Fig. 54, from Denmark, Madsen, Heft iii.—Fig. 61, from near Mainz, Lindenschmit, Taf. iii.—Fig. 62, from the Museum at Wiesbaden, Lindenschmit, Taf. iii.—Fig. 63, from Altona, in Courland; this form has some affinity to Class G, but is introduced here on account of the expansion of the blade.—Figs. 64, 65, and 66, from Italy, in the British Museum, the metal of the shaft slightly sunk to produce a stop.—Fig. 67, from Fiesole, Italy, the metal part of the shaft further reduced.—Fig. 68, from Baron von Stackelberg’s collection, in the British Museum, also described in Klemm, Werkzeuge und Waffen, p. 103, Fig. 180; said to be from Greece, but its close resemblance to those from Italy is remarkable.

Class F.—The same form as Class E, but having the flanges bent by hammering over the stop; the flanges appear to have been cast upright, as in Class E, and to have been bent over the cleft handle after hafting; by this means the necessity for binding the blade on with thongs was obviated. This class forms a transition to the socket type.—Figs. 69, 70, 71, from Ireland, in my collection.—Fig. 72, from the Royal Irish Academy collection, having a loop on the side. See Catalogue R. I. A., ‘Bronze,’ page 379. The introduction of the loop appears to be synchronous with the abandonment of the binding, the overlapping flanges answering that purpose by enclosing the bent portion of the handle, and requiring only that it should be fastened by the loop to prevent its falling off the end of the handle.—Fig. 73, from Denmark, in my collection.—Figs. 74, 75, from Denmark, Madsen, Heft iii.—Fig. 76, from the Museum at Hanover, Lindenschmit.—Fig. 77, from the Museum at Munich, Lindenschmit, Taf. iv.—Fig. 78, from Möringen, Switzerland, Keller, Plate xli.—Fig. 79, from Nidau-Steinberg, Switzerland, Keller, Plate xxxv.—Fig. 80, from Hallstatt; Von Sacken.—Fig. 81, from Italy, British Museum.

Class G.—The pocket type. The bent portion of the handle in this ease was retained in its place by pockets cast on each side of the shaft of the blade; it seems doubtful whether this, or Class F, is to be regarded as the nearest approach to the socket type. In Class F the overlapping was produced by hammering the metal; but Class G is a further advance in the casting process.—Figs. 82 and 83, from Ireland, in my collection; the latter with loop; the pockets or pouches to receive the points of the bent handle are shown in the sections.—Fig. 84, from France; see Matériaux pour l’Histoire de l’Homme.—Fig. 85, found twelve leagues south of Oviedo, Spain, in the collection of the Society of Antiquaries.—Fig. 86, from Andalusia, Spain, British Museum.—Fig. 87, from Denmark, Madsen, Heft iii.—Fig. 88, from the collection at Munich, Lindenschmit.—Fig. 89, from the collection at Hanover, Lindenschmit.—Fig. 89 a, an iron celt of the same form, still in use by the Kalmucs, Siberia, Prehistoric Times, p. 26.

Class H.—The socket type. In some of the specimens of Class G, as for example Figs. 82 and 83, the metal portion of the shaft of the blade dividing the two pouches is reduced to a minimum. The next step was to do away with it altogether and enlarge the sides of the pouches so as to form a single socket. By this means the bent handle no longer required to be cleft to receive the blade, but was inserted whole into the socket, producing greater firmness, each blow of the axe serving to fix it more securely to its handle. The loops, seen only occasionally on Classes F and G, are almost invariably present in Class H.—Figs. 90, 91, 92, 93, 94. Socket celts of bronze, from Ireland and England, in my collection; the form with square sides is very uncommon in Ireland; in Fig. 92 a representation of the overlapping flange of Class F is cast on the surface of the socket.—Fig. 94a, a socket celt of wrought iron with loop, from Merionethshire, British Museum; Archaeologia Cambrensis, vol. i, third series, p. 250.—Figs. 95 and 96, the same forms from France. See Matériaux, &c. The square-sided celt is common in the north of France.—Fig. 97, from Alemquez, Portugal; Coll. Société des Archit. Portugais.—Fig. 98, from Denmark, in my collection.—Figs. 99, 100, Denmark, Madsen, Heft i.—Fig. 100 a, an iron socket celt, from the moss of Nydam, in Slesvik, of the iron period; Engelhardt, Denmark in the Early Iron Age (1866), Pl. xv; believed, from the Roman coins found with it, to be of the third century A.D.[217]—Fig. 101, from the collection at Hanover, Lindenschmit.—Fig. 102, from the Museum at Mainz, Lindenschmit.—Fig. 103, socket celt of iron, from Golssen, Klemm, Fig. 195.—Fig. 104, socket celt of iron, from Thuringia, Klemm, Fig. 194.—Fig. 105, of bronze, from Unter Uhldingen, Switzerland; Keller, Pl. xxix.—Fig. 106, of iron, found near Marin, Switzerland, the socket formed by beating over the blade on one side only; the socket is not quite completed; see Keller, Pl. lxxi.—Fig. 107, the same form of iron, found near Marin; the socket is closed and completed all round, Keller, Pl. lxxi. These specimens in iron may be regarded as connecting links between Classes F and H. Viewing the occurrence of iron celts of this form, it appears not impossible that the introduction of the socket type and the sudden abolition of the central division may have been suggested by the use of the more malleable metal, by means of which the fabricator acquired the art of forming a socket by bending over the metal on one side; the inutility of the central division would thus become apparent.—Fig. 108, bronze socket celt with loop, from Hallstatt, Von Sacken.—Fig. 109, exactly the same form in iron, from Hallstatt; a portion of the wooden handle is still shown in this specimen.—Figs. 110 and 111, bronze socket celts, from Italy, of a variety peculiar to that country, British Museum.—Fig. 112, socket celt of copper, from Hungary, believed by the author to be the only known specimen of pure copper; Keller, Pl. lxxviii.—Fig. 113, bronze socket celt, from Hungary, British Museum.—Fig. 114, bronze socket celt, with two loops, from Kertch, British Museum.—Fig. 115, bronze socket celt, from the province of Viatka, Russia. See Matériaux, &c.—Fig. 116, bronze socket celt with two loops, from the Ural, Russia.—Fig. 117, mode of hafting, Classes A, B, and C.—Fig. 118, mode of hafting, Classes D, E, F, and G.—Fig. 119, mode of hafting, Class H.

In a paper lately read to the Society of Antiquaries by Dr. Thurnam,[218] he has drawn attention to the fact that none but celts of the most primitive type, viz. those belonging to Classes B, C, D, and the most rudimentary form of Class E, have been found in the British tumuli. Scarcely a single instance of the more developed palstave or of the socketed celt has as yet been discovered; the only exceptions being a bronze socket celt found in a tumulus on Plumpton Plain, near Lewes, and a diminutive bronze socket celt found in a tumulus at Arras in the Yorkshire wolds. These Arras barrows are known, however, to belong to the iron age; having produced, amongst other articles composed of that metal, the iron tire of the wheel, and trappings of a war chariot. We learn from this that the discoveries in the tumuli confirm in point of time the order of development inferred from a consideration of the implements themselves.

From the foregoing detailed description of Plate XVIII we are enabled to draw the following conclusions, viz.:—(1) That in each of the divisions of Europe therein represented, traces of the development of the celt, from its simplest to its most complex form, have been discovered; the earliest forms being in imitation of those of stone, and being not unfrequently constructed of pure copper. Where some of the connecting links are wanting in the table there is reason to suppose the absence of those links may be the result of imperfect information, and does not necessarily imply a flaw in the continuity of development. (2) That, notwithstanding the simultaneous development which appears to have taken place in different countries, we may nevertheless observe slight differences in the details of construction, which are sufficient to give a distinctive character to the celts of each separate region. Thus, for instance, the celts from Ireland are, as a general rule, shorter and less elegant in form than those found on the Continent. Class C, consisting of stop celts without wings, though common in Great Britain and Ireland, is, so far as I have been able to ascertain, unknown on the Continent. On the other hand, Class D, having wings without stops, is rare in Ireland, but common in France, Denmark, Germany, and Switzerland. The development of this class of celt into a nearly circular edge, as represented in Figs. 34 and 35, is peculiar to the south of France, though traces of it are observable in the celts from Germany, Fig. 40. Class E, having both stop and flange, is found in a more rudimentary stage in Ireland than elsewhere. The palstaves of this form, having shoulders on the side of the blade, are peculiar to Italy and Switzerland, Figs. 66, 67, and 68. Class F, with overlapping wings, is but slightly developed in Ireland, but is fully so in Italy, Germany, and at Hallstatt. Class G, the double pocket variety, has its head quarters in the north-west of France, but is also known in Ireland, Denmark, Spain, and Germany; it is, in so far as I have been able to ascertain, unknown in Italy. Class H, the socket type, varies greatly in different countries; the square form, Figs. 93, 94, 95, 96, 100, and 102, is exceedingly rare in Ireland, but common in France. The socket celts from Italy, Figs. 110 and 111, are of peculiar type, and evidently derive their form from the winged palstave of the same country, Fig. 67. Socket celts of iron have been found at Hallstatt, and in Switzerland, Denmark, Germany, and North Wales. The representation of the overlapping wings, cast on the surface of the socket celt, Figs. 92 and 101, is common in England and Germany, but exceedingly rare in Ireland. The double-looped socket celt, Figs. 97, 114, and 116, appears to be especially characteristic of the Eastern provinces of Russia and Siberia, though found occasionally elsewhere.

In attempting to account for the varieties, which I have described, in the details of construction, coupled with a general uniformity of design throughout the entire region of distribution of these weapons, we may, I think, draw an exact parallel between the development of bronze celts and the development of the forms of cannon between the fourteenth and the nineteenth centuries. From Europe to China we know that the form of cannon has developed upon the same plan. In the same way that the overlapping wings of the palstave were represented on the faces of the socket celt, so the rings of metal which bound together the bars of which the ancient bombard was composed, were represented on the surface of the cast bronze cannon which superseded it. In every country the general type of development of cannon has been the same, but the details of construction have varied in each. Even in our own time, the introduction of breech-loaders has been synchronous throughout Europe; but the French and English cannon are not perfectly identical. Now, the cause of this is sufficiently well known. There has been constant intercommunication between the several countries throughout the whole period of the development of this weapon. Each new improvement as it occurred has been communicated from one country to another, either by contact in war, or by peaceful intercourse; but each country has fabricated its own weapons, and has by that means contrived to give them a national character.