54

CELLULOID

CELSIUS

first made by Mr A. Parkes of Birmingham in 1855 or 1856. It chiefly consists of a dried solution of gun-cotton (pyroxylin), or of what is nearly the same thing, and oil. A variety of it can be made with pyroxylin and camphor. It resembles ivory, horn, tortoiseshell, and hardened india-rubber, as regards certain properties.

The pyroxylin is prepared by treating Cellulose (q.v.) from such vegetable materials as cotton or flax waste, rags, paper-makers' half-stuff, or paper | itself, with a mixture of one part of strong nitric acid and four parts of strong sulphuric acid. It is convenient to call the product so obtained pyroxylin, although the two things are not quite identical. The distillate obtained by distilling wood naphtha with chloride of lime is used as a solvent for the pyroxylin, but other solvents, such as nitro-1 benzol or aniline, and some camphor are added with advantage. When the excess of solvent is removed from the pyroxylin, it is mixed with a considerable quantity of castor-oil or cotton-seed oil, and made into a dough or paste between heated rollers. For a hard compound the quantity of oil should be less than the pyroxylin, for a soft one it should be greater. Chloride of sulphur is sometimes added to the oil. When articles made of celluloid are in a partially manufactured state, they are soaked in bisulphide of carbon or chloride of lime to remove any trace of solvent, which would render them apt to shrink if allowed to remain. Celluloid is of a somewhat combustible nature unless the substances used to colour it are such as will neutralise this, or unless some non-combustible chemical, tungstate of soda for example, is added to it.

Properties and Uses.-Celluloid has many valuable properties. It is buff or pale brown in colour, but it can be made as white as ivory, which it much resembles, or manufactured in a transparent state. It can be moulded or pressed into any form, and turned, planed, or carved. Neither the atmosphere nor water affects it. It is elastic and can be united by its own cement. In a plastic condition celluloid can be spread on textile fabrics, or it may be made as hard as ivory, for which it is largely used as a substitute. Billiard-balls, piano-keys, and combs are made of it, the latter two articles extensively. It can be coloured to represent amber, tortoiseshell, or malachite. In imitation of red coral it has been a good deal used for jewelry. Like vulcanite, which it excels in durability but exceeds in price, it has very numerous applications. We need only mention brush-backs, knife-handles, buttons, napkin-rings, card-cases, thimbles, and dolls. useful for optical instruments and for some surgical instruments. One of the most recent applications of it is for shirt fronts and collars. The manufacture of celluloid, although an English invention, has been most largely developed in the United States, where it is mostly, if not entirely, made by one firm, the Celluloid Manufacturing Co., Newark, New Jersey,

who use this word as a trade-mark.

It is

becomes much more considerable. Cellulose has the chemical composition CHO, and spec. grav. 1·52. Among its familiar natural modifications gum is an isomer, and starch-dextrin and grape-sugar are all of similar ultimate composition, while its woody and corky modifications (lignin and suberin) possess stains cellulose yellow or brown, but blue when an increasing proportion of carbon. Iodine alone strong sulphuric acid has been previously added. mersion in the cold converts it into a tough and dense Strong hot sulphuric acid chars it, while brief im. prolonged treatment dissolves it altogether. Dexmodification, well known in parchment paper, and trin may thus be prepared and next transmuted, by boiling the watery solution, into grape-sugar (see DEXTRINE, GLUCOSE). By immersion in a obtain Gun-cotton (q.v.), while dilute nitric acid mixture of strong nitric and sulphuric acid we or potash oxidises it into oxalic acid. Ammoniacal oxide of copper dissolves it without change, as is shown by its reprecipitation on dilution. By heating in closed vessels under pressure a dense coallike mass is formed, while in ordinary dry distillation, gas, tar, and acetic acid are given off, proin nature and on the chemistry of gas-making. In cesses which throw light on the formation of coal natural decomposition cellulose turns yellow and brown with gradual formation of humus. See

SOILS.

Although so constant and characteristic a proits formation are still very obscure. From that cellduct of vegetable life, the conditions and mode of cellulose-walled state and an active and wall-less cycle or rhythm of change between the passive and one, which is so characteristic of the lowest forms of life, and of which we find surviving traces (e.g. the rejuvenescence of the pollen-grain) in the reproductive processes of even the highest plants (see CELL), it would appear that there is some relation between this increased passivity and the formation of cellulose. And in this way arises the speculation that cellulose may be viewed essentially as a (mechanically coherent and thus useful) excretion, an incompletely utilised waste product corresponding to the carbonic acid and water given off by the completer respiratory oxidation and larger Once formed by the plant, it may be again absorbed, as evolution of energy of the active phase. is well seen in the union of a row of cells into a continuous vessel, or in the consumption of endosperm of a seed during germination. Many seeds, such portion of their reserve material in this form; and as vegetable-ivory or date, have a great prothis must be digested into glucose by the growing embryo, and again worked up into new protoplasm, which deposits cellulose as before. Like the plant itself, the similar digestive ferments of the animal might thus be naturally expected to digest cellulose; and this is actually, to some extent, the case with the delicate young cell-walls of many green vegetables, as can be experimentally verified, even in man; while in herbivorous animals this power is much developed, and the nutritive utilisation of their fodder is thus increased to an important

extent.

The cysts of amoebae and other protozoa appear to be at least largely composed of cellulose, and the external tunic of ascidians (see TUNICATA) is of identical, or at least isomeric, composition. Cellulose has been described as a pathological product, even in brain-tissue; and Chitin (q.v.), a very char acteristic and in many respects comparable animal product, has been sometimes viewed as cellulose in association with a proteid substance.

CELSUS

| Magnus Celsius (1621-79), a professor of Astronomy and decipherer of the Helsing runes, and the nephew of Olof Celsius (1670-1756), professor of Theology at pala, author of the Hierobotanicon, and an early friend and patron of the great Linnæus. Anders became in 1730 professor of Astronomy at Upsala. Two years later he set out on a scientific tour, visiting the observatories of Nuremberg, Rome, and Paris. After his return he published his De barrcationibus pro figura telluris determinanda in frx' a habitis (Upsala, 1738). In 1740 he had the satisfaction of seeing a splendid observatory erected at Upsala, and here he laboured till his death, 25th April 1744. The Inscriptions of the Swedish Academy contain many papers by Celsius on astrotomy and physics. It is, however, as the first constructor ; 1742) of the thermometer now chiefly sel by scientific men, that he is best known. In at the space between the freezing point and the testing point of water is divided into one hundred saces, hence Celsius's thermometer is often called the centigrade or centesimal scale. See THER

MOMETER

Celsus, a Platonic philosopher, but tinged with ¡Escureanism, who lived in the 2d century after Grist, was a friend of Lucian, and wrote, about 176 180 during the persecution of Marcus Aurelius, under the title Logos Alithes (true word '), the first notable polemic against Christianity. The book rself has perished; but considerable fragments have been preserved as quotations given by Origen in Lis answer, Contra Celsum, in eight books. In the fragments which are very interesting, as showing ! the views of a heathen philosopher in regard to Cristianity-Celsus, with great acuteness and wit, at without depth or earnestness of thought, prefers against the new religion charges of unphilosophi ca ness and blind credulity; and especially endeavours to convict Christians of self-contradiction in their spiritual doctrine contrasted with their athropomorphic representations of Deity; in their religious arrogance contrasted with their confession of wintuiness; and in their views of the necessity et reiemption. He also reproaches Christians with eir party divisions and ever-varying opinion, and enles them as worms in a corner who think - occupy the centre of the world. Celsus is that the Supreme God can have no contact with the material world, the creation of which is the work of inferior deities or demons. He regards evil as an essential property of the material world; he way: There neither has been in former times, is there now, nor ever shall be, an increase or diminution of evil. The nature of the universe is ever dentical, and the production of evil is not a variable quantity.... It is evident that those who sin by nature and by habit cannot be changed in any respect either by punishment or by pardon." He charges Christians with having remodelled | · The Gospel" from the "first writing" three times, far times, and many times.' However, as Origen Tarked, almost everything of an historical kind to which Celsus refers is to be found in our Gospels,

ially the Synoptics. See Keim, Celsus' Wahres # 1873); Aube, La Polémique Paienne in Les Ferrations de l'Eglise (1878); Pelagaud, Etude ver Cefar (1878); Froude's Short Studies, vol. iv. ; the article on ORIGEN.

Celsus, AULUS CORNELIUS, a Latin physician writer, who probably flourished about 50 A.D., a: wrote not only on medicine, but also on rhetoric, history, philosophy, the art of war, and agriculture. His style is succinct and clear, but full of Græcisms, The only great work of his which survives is the In Medicina. The portions relating to surgery are exceedingly valuable, as giving an account of opanions and observations of the Alexandrian

school of medicine. Indeed, to Celsus, next to Hippocrates and Galen, we mainly owe our knowledge of the medicine of antiquity. Celsus's works were translated into English in 1756. Next to the first edition (1478) the most important are those of Targa (1769) and Daremberg (1859). See Broca's Conferences historiques (1865).

which the axe-heads of the early inhabitants of Celt (Lat. celtis (?), a chisel'), a name by Europe are known among British and French archaeologists. The Scandinavian archaeologists use the word 'axe' and not 'celt.' Its use is now solete. The word is generally believed to have considered pedantic, and it is fast becoming oboriginated from a misreading of Job, xix. 24, in the Vulgate, where Celte, understood to mean with a chisel,' was read in place of Certe, verily' (cor responding to the 'for ever' of the English Bible). Celte or Celtis is not elsewhere found in Latin. Notes and Queries (1878), vol. ix. p. 463; vol. x. p.

73.

See

Celts are either of stone or of bronze. Stone celts vary in length from about 1 inch to 22 inches; but the most common size is from 6 to 8 inches in length, and from 2 to 3 inches in breadth. They are made of almost every kind of stone, and show considerable diversity of shape, almost all, however, having more or less resemblance to the mussel-shell. The ruder celts are generally of slate, shale, schist, or grit; the finer, of flint, porphyry, greenstone, syenite, or agate. Many of the finer celts are beautifully shaped and highly polished. Some very remarkable examples of this class are in the National Museum of Antiquities, Edinburgh, and one found near St Andrews, in Scotland, is described by Sir David Brewster in the Philoso phical Journal for 1823. The stone celt was fastened into a handle of horn, bone, or wood.

Bronze celts vary in length from about 1 inch to 8 or 10 inches, the most common length being about 6 inches. They show much greater diversity of shape than the stone celt. As many as four classes have been distinguished by archæologists: (1) The flat wedge-shaped celt, most nearly resem bling the common form of the stone celt. (2) The flanged celt, with the side edges more or less overlapping, and a stop-ridge or elevation between the blade and the part which received the handle. (3) The flanged-edges celt, with side greatly overlapping, with or without the stop-ridge, but with a loop or ear upon one side. (4) The socketed celt, or the celt with a hollow to receive the handle, and generally with a loop or ear upon one side. They are sometimes ornamented with raised lines or circles formed in the mould in which they were

cast.

Both stone and bronze celts were probably used for many purposes, serving for chisels, adzes, and axes, as well as for weapons of war, like the stone hatchets of the South Sea Islanders. See under FLINT, STONE AGE, and BRONZE AGE.

Celtibe'ri, a brave and powerful people of ancient Spain, supposed to have sprung from a blending of the aboriginal Iberians with Celtic invaders from Gaul. They inhabited a large inland district of the peninsula, corresponding to the south-west half of Aragon, nearly the whole of Cuença and Soria, and a great part of Burgos, but the name Celtiberia had often a wider signification, including the country as far south as the sources of the Guadalquivir. The Celtiberi were divided into four tribes, the chief the Arevacæ and Lusones, and were unquestionably one of the bravest and noblest peoples in the peninsula. Their cavalry and infantry were equally excellent. For many years they withstood the efforts of the Romans to subdue them, and it was not till after

the death of Sertorius (72 B.C.) that they began to adopt the Roman language, dress, and manners. The chief cities were Legobriga, the capital; Bilbilis, the birthplace of Martial; and Numantia, destroyed by Scipio Africanus after a desperate ten years' resistance, 133 B.C.

Celtic Ornament, a peculiar development of the system of iron-age decoration prevalent in the British Isles. Its history is divided into two periods by the introduction of Christianity, which engrafted on the older style a number of new elements of decoration brought into the country with the manuscripts of the gospels and psalters, and supplied new forms for the display of these elements, such as churches and crosses, shrines, bells, and crosiers. In its pre-Christian stages, ranging approximately from two or three centuries before the Christian era to about the end of the 6th century A.D., it appears principally in connection with the metal mountings of harness and horsetrappings, and on shields, sword-sheaths, mirrors, armlets, and other articles of personal use and ornament. The material is usually bronze, but occasionally silver or gold. The principal characteristics of the pre-Christian style are its preference for elliptical curves and divergent spirals; its use of chased or engraved lines or dots as a diaper in the spaces of the general design in contrast with other spaces left plain; its use of repoussé work, sometimes in very high relief, at other times in low relief on thin plates riveted on in their places in the general design; the production of peculiar patterns often in excessively high relief in the casting; and the employment of champ-levé enamels of red, yellow, blue, and green, and settings of coloured vitreous pastes. One of the finest examples of such settings occurs in the decoration of an oval shield of bronze, from the bed of the Thames, ornamented with Celtic patterns in relief, enriched by twenty-seven settings of red enamel, kept in their places by small cruciform ornaments of bronze riveted in the centre of each. There are to be seen in the National Museums of London, Edinburgh, and Dublin enamelled shields, sword-sheaths, and ornaments of horse-trappings in bronze, of great beauty and excellence both of design and workmanship, and other articles in bronze, silver, or gold, ornamented in repoussé work or in relief, with or without enamel as an enrichment, found in many parts of England, Scotland, and Ireland, in pagan grave-mounds, in crannogs or lake-dwellings, in earth-houses, in the beds of lakes and rivers, or in casual deposits under the soil for concealment. In a work entitled Hora Ferales, Mr Franks of the British Museum has figured in colours many of the, best of these remarkable products of the earliest known process of champ-leve enamelling, and adduced evidence to show that it and this peculiar style of Celtic ornament which accompanies it were of indigenous origin, and at this early period peculiar to the British Isles.

The re

markable development of Celtic ornament which succeeded the introduction of Christianity was characterised by the association of interlaced work and fretwork with the elliptical curves and divergent, spirals which up to that time had been the principal elements of Celtic design. To these were occasionally added a step-like pattern, and diapers of the Z and I shaped patterns sometimes seen in Chinese decoration. The interlaced work was elaborated with excessive care into patterns, presenting an infinite variety of combinations pleasing to the eye, and capable of being harmoniously treated in colours. It was sometimes a simple ribbon-like band, which might be plain, or divided in the middle, or divided into three by lines close to the margin; or the inter

CELTS

lacing band might be replaced by an elongated animal form with its feet, its tail, and its top-knot drawn out to interlace with each other, and with the corresponding parts of other lacertine forms, the whole forming a diaper of quaintly expressed and complicated construction. The fretwork was also elaborated with much ingenuity into most complicated patterns, a special feature of the style being its partiality for diagonal frets and patterns produced by combinations of oblique lines, in direct contrast to the fretwork of Greek and Roman art, which was essentially rectangular. The elliptical curves and divergent spirals of the older style, which had received their only expression in the solid forms proper to metal-work, were found to be equally capable of adaptation to the purposes of the illuminator, and by a similar process of combination and elaboration they also produced patterns and diapers of inexhaustible variety and beauty. A special feature of Celtic decoration was its tendency to divide the surface to be decorated into a series of panels, each of which was treated as a separate whole. The finest examples of Celtic ornament are unquestionably to be found in the grandly illuminated pages of manuscript copies of the Gospels, from the 7th to the 9th century. Of these the most famous for the elaborate nature of their ornament and the beauty of their colouring are the Book of Kells in Trinity College, Dublin, and the Lindisfarne Gospels in the British Museum. Of enamelled metal-work in this period there may be mentioned the Ardagh Chalice, perhaps the most elaborate and beautiful of all the products of Celtic art, the Lismore Crosier, and the Monymusk Shrine. Examples of filigree-work, and chasing or engraving in gold and silver of the highest excellence are found in the Tara Brooch, the Ardagh Brooches, the Rogart Brooches, and the Hunterston Brooch, the Shrine of St Patrick's Bell, the Shrine of St Manchan, and the Cross of Cong. The approximate dates of the metal-work of the highest excellence range from the 10th to the 12th century. For sculpture in stone it is only necessary to refer generally to the incised slabs and sculptured crosses of Scotland and Ireland, ranging from the 9th to the 12th centuries, the special characteristics of their decoration being the same as those of the manuscripts and metal-work already mentioned. For illustrations, see BROOCH, CROSS, SCULPTURED STONES. See further Kemble's Hora Ferales, edited by Latham and Franks (1863); Anderson's Scotland in Early Christian and Pagan Times (1881-83); Westwood's Palæographia Sacra Pictoria (1845), and Fac-similes of the Miniatures and Ornaments of Anglo-Saxon and Irish Manuscripts (1868); O'Neill's Fine Arts of Ancient Ireland (1863), and Sculptured Crosses of Ancient Ireland (1857); Stuart's Sculptured Stones of Scotland (Spalding Club, 1856 and 1867); and Miss Stokes's Early Christian Art in Ireland (1887).

Celtis. See NETTLE-TREE.

Celts. The Celtic nations of antiquity had no comprehensive name. Those of the Continent were called Galli by the Romans, and less usually Celta. The Greek equivalents for these terms were Galatai or Galatæ, and Keltoi or Celti. But neither Greeks nor Romans regarded the British Isles as belonging to the Celtic world. They were situated outside it, and lay over against it in the sea; still it was known to men like Julius Cæsar that certain portions of Britain were in ited by Celts in the sense of Galli or Belgae.

Celtic ethnology involves and we shall speak of them according to the more palpab

'ifficult questions, article mostly Fons of speech;

and in order to proceed as much as possible from the known to the unknown, we begin by classifying their idioms. These, whether dead or still spoken, belong to the Aryan or Indo-European family of Languages, and those of them spoken in modern times divide themselves into two groups-viz. Goidelic and Brythonic. (1) The Goidelic group embraces the dialects termed Gaelic, that is to say, Irish Gaelic, or Irish as it is now more frequently and briefly called; Manx Gaelic, or the Gaelic dialect not yet extinct in the Isle of Man; and Scotch Gaelic, or the Gaelic spoken in the Highlands and Islands of Scotland. In ordinary Scotch and English parlance this is what is understood by the word Gaelic when it is used without any qualification. In order to resist one of the delusions to which charlatans are always leading the unwary, it is right to say that the words Gael and Gaelic have nothing to do with Galli. Gael is the simplified English spelling of a word which is now written in Scotch and Irish Gaelic Gaidheal, with an evanescent dh; but the most ancient form known of it was Goidel, whence the adjective Goidelic, which has been resorted to by Celtic scholars as ap; icable equally to all three Gaelic subdivisions of the Celtic group here in question. The Celtic languages of this group are sometimes also called Erse, which is a term derived from the Scotch form of the adjective Irish; this was Ersch or Yrisch, the longer and shorter forms of which appear, used without any distinction, by Kennedy in his answer to the poet Dunbar, when the latter had called Kennely an Ersch brybour baird' and an Ersch Katherine, in reference to his alleged extracto from the Irish Scots of Galloway and Carrick. Kennedy's reply contains the following line (see Murray's Dialect of the Southern Counties of Scot land, 1873, pp. 43-44):

Thou luvis nane Erische, elf, I undirstand, and he goes on to add

Thy fore fader maid Ersche and Erschmen thin. 12) The Brythonic group embraces the following languages: Welsh, Breton, and Cornish, which has been extinct now for about a century. Two of | these belong to Great Britain, and one, the Breton or Armoric, to Little Britain on the other side of the English Channel. These three might be colectively termed British or Britannic, but that both these adjectives have connotations which would be misleading, as they tend to confusion; so here, also, a neutral form, Brythonic, is used, which is derived from Brython, one of the Welsh wds for the Welsh and the so-called Ancient Britons, whence their language is sometimes called Bruthoneg in Welsh. This last was in Cornish Brethomer, and in Breton Brézonek, meaning respectively the Celtic of Cornwall and of Brittany, Erython or Britto was the national name of all peoples of this branch, just as Goidel or Gael may treated as the national name of the other branch.

All this applies only to the neo-Celtic nations, or the among whom Celtic languages are or have been in use in modern times, and a question of hgreater difficulty presents itself when one attempts to classify likewise the continental Celts at ar ient history. The reason for this is chiefly, the fact that the linguistic data become more preParris as one goes back. Thus, for example, the me of the ruling people of ancient Gaul has he left us only in a very few inscriptions, so at our knowledge of it from that source has to be rumplemented by the study of Ganlish proper hates of which a considerable number is extant m Latin inscriptions and in the writings of Roman and Greek authors. Now, when we apply the

57

test of some of the most palpable differences that are known to exist between the Goidelic and the Brythonic idioms to the remains of the Gaulish language, we find at once that it is to be ranked with the Brythonic dialects, and not with the Goidelic ones, and our Brythonic group becomes what may be more exactly described as a GalloBrythonic one. This further suggests the question whether there was no continental Celtic idiom which partook of the characteristics of the Goidelic branch. The probability is that there was; for one finds Sulpicius Severus, an ecclesiastical writer of the 4th century, distinguishing between Celtic and Gallic or Gaulish, as if both were spoken in his time. (See Dialogue i. 26, in Migne's Patr. Lat. vol. xx. col. 201: Tu vero, inquit Postumianus, vel Celtice, aut, si mavis, Gallice loquere, dummodo jam Martinum loquaris.') And the use of the two names Celta and Galli would seem to point to the same inference-viz. the exist ence in Gaul of two Celtic peoples, the one, probably, superimposed on the other, as on a vanquished population, or driving it towards the south and west. Thus, if the Celtic language which Sulpicius Severus distinguished from Gaulish should be ranked with the Goidelic dialects, we should have alongside of a Gallo-Brythonic group another which might be called Celto-Brythonic were it not inconvenient to use the words Celt and Celtic in two senses. For while the modern usage applies them indifferently to the whole family, Sulpicius indicates a narrower sense; and so, in fact, had Cesar done centuries before, when he wrote that one of the three peoples of Gaul was called Celta in their own tongue. He states that these Celtæ proper, so to say, were separated by the Garonne from the Aquitani, and by the Seine and the Marne from the Belge. In other words, their country extended from the Garonne to the Seine and Marne, and other Roman writers give it the name of Celtica; and Dionysius of Halicarnassus had heard of a river Celtus, from which Celtica was supposed to derive its name. From this narrower Celtica, in the sense which Roman writers gave it, one might form the adjective Celtican, to apply to its people, in order to avoid the confusion which must arise from calling them Celts, whilst using that word also of the whole family.

In order to show the philological reasons for this classification, it would be necessary to go into a variety of details; but let one of these suffice for the present. The Gallo-Brythonie dialects used p where the others would have qu. Take, for example, the early inscriptional Irish for the genitive of the word for son'; it was maqri, corresponding to a nominative which appears as mace or mac in the oldest manuscript Irish; and mae is still the word for boy or son' in all the Goidelic dialects. Now the early Brythonic form of this genitive would have been mapi, while in the oldest manuscript Welsh we have map, and in later Welsh mab, boy' or 'son. From this word was formed another, mabon, a 'boy' or 'youth;' and this in its old form appears in Latin inscriptions as maponus in Roman inscriptions found in Britain in honour of the Celtic god Apollo Maponus, so called in reference to his youthfulness, Now from Gaul we have such names as Eporedorix, Parisu, Petrocorii, and many others, with the consonant p; but every now and then we have also names with qu, such as Sequana and Aquitani, together with several instances from Spain, where a people of the same Celtic branch as those of Celtica had also probably established themselves.

So far, then, as one can get philological data to reason upon, it would seem that the west of Europe had in early times been subjected to two Celtic

invasions; the one is represented by the Celts whose position, geographically speaking, is the farthest from the home of the Aryans. These would be the Celticans of Gaul and Spain, as compared with the Gallic tribes to the east of them towards the Rhine and the Alps; the same relative position is also taken up by the Goidelic Celts of the British Islands, occupying, as we find them doing, the Isle of Man, Ireland, and the Scotch Highlands and Islands. The other, here represented by the Brythons, must have come later and driven out the Goidels, or subdued them, in the rest of this island. This may be supposed, also, to have been the case on the Continent, so that we have to regard the later comers, the Galli, as invaders and conquerors forming another Celtic population. In the eastern portions of Gaul they may have formed the bulk of the population, but in the rest of that country they probably only constituted a ruling class of comparatively small importance in point of numbers. Such a state of things would adequately explain the great dearth of linguistic remains belonging to the older and subjugated people. Roman authors and other strangers would naturally speak most of the ruling classes, and information about the others must reach strangers through the medium of the Gallic rulers and their language, at any rate, so far as concerns the time before Latin became the official tongue of all Gaul. A somewhat similar conclusion has been arrived at by studying the burials and megalithic monuments of France and the neighbouring lands to the east of it. In Central and Western France menhirs, dolmens, and cromlechs prevail, while the eastern side of France shows the prevalence of mounds and barrows, which are here and there found penetrating into the other domain, giving us a sort of rude sketch, as it were, of an invasion advancing irregularly towards the west. See M. Bertrand's Archéologie Celtique et Gauloise; also K. von Becker's Versuch einer Losung der Celtenfrage (1883), pp. 114-119.

For reasons already indicated, the question of Celtic ethnology is a very difficult one, but it is considerably more difficult than would appear from what has here been mentioned; for besides two Celtic sets of invaders, there are also to be taken into account the non-Aryan races that previously occupied the countries to which the Celts came. These pre-Celtic populations probably survived in considerable numbers, and one of the effects of a second Celtic invasion may be supposed to have been to force the earlier Celtic settlers to amalgamate with the ancient inhabitants, and to make common cause with them against the later Aryan hordes. So it may be expected that the language of the Goidelic Celts will prove to have absorbed a larger non-Aryan element than that of the Brythons. Similarly, one might take for granted that the physical type of the people speaking the Goidelic dialects should prove less purely Aryan; but this feature is obscured by the fact of the counter-invasions which Wales and other western portions of Britain have undergone in historical times at the hands of Ireland. Lastly, it is right to add that in so far as the people, whose language is or has been Celtic, are Aryans, one might expect the type to be that of tall men, with more or less light hair and blue eyes; on the other hand, the smaller men, with dark hair and black eyes, which it was the fashion till lately to regard as the genuine and typical Celts, are probably not to be regarded as Celts at all, but as Ivernians or representatives of the pre-Celtic and non-Aryan race, whose hunting-ground the soil of the British Islands may be said to have been long before the first Aryan set foot in them.

The Celtic languages and literatures will be found

CEMENTS

under BRITTANY, CORNWALL, GAELIC, IRELAND, WALES. See also ARYAN RACE AND LANGUAGES, ETHNOLOGY, PHILOLOGY, DRUIDISM.

Besides the works already mentioned, the following should be consulted: Müllenhoff's Deutsche Altertumskunde (Berlin, 1887); Windisch's article 'Keltische Sprachen' in the Allgemeine Encyklopedie der Wissenschaften und Künste, together with the reviews on the same in the Revue Celtique, vol. vi. pp. 395-400; Hübner's Inscriptiones Britannia Christiana (Berlin, 1876); Brambach's Corpus Inscrip. Rhenanarum; and the volumes of the Corpus Inscrip. Latinarum, published by the Berlin Academy, especially those for Britain (vii.), Spain (ii), Gallia Narbonensis (xii.), Gallia Cisalpina (v.), and Illyricum (iii.).

Cements. These may be roughly divided into three classes: (1) The stone cements, including Roman and Portland cements, and ordinary mortar, which are used in thickish layers for uniting stone and brick work, and for protective coverings to buildings; (2) substances which form binding joints of much less but still appreciable thickness, such as white lead, red lead, and putty; and (3) cements which require to be used in extremely thin coatings, such as glue, isinglass, and dissolved caoutchouc.

Ordinary Mortar is a mixture of slaked lime (calcium hydroxide) and sand, made into a paste with water. Generally one part of lime to three or four parts of sand are used, but the proportions vary according to the purity of the lime employed. Very pure or fat lime, such as that made by burning white chalk or white marble does not make so good a mortar as lime obtained from less pure limestones, which are by far the most abundant. The more thoroughly the ingredients are intermixed, the more complete will be the subsequent hardening of the mortar. As commonly laid in the joints of brick or stone work, mortar sets sufficiently fast to allow building operations to proceed from day to day with occasional longer intervals, but it takes years-perhaps in many cases centuries-to reach its maximum hardness. setting and subsequent slow hardening of mortar are usually considered to be due, in the first instance, simply to the loss of water, and afterwards to the absorption by the lime of carbonic acid from the atmosphere, the carbonate of lime thus formed binding together the sand and stone. It is doubtful, however, if this is an altogether satisfactory explanation. The mortar used in many medieval buildings is largely mixed with small pebbles. In a number of cases this has proved to be of a more durable nature than the stone used along with it.

The

Puzzolana or Pozzuolana, a loosely coherent volcanic sand found at Pozzuoli, near Naples, has been long celebrated for its property of forming a hydraulic cement when mixed with ordinary lime. It is composed of silica, with a little magnesia and potash or soda, alumina, lime, and oxide of iron.

Roman Cement.-Certain natural mixtures of lime and clay are called cement-stones. The clays of some of the newer geological formations in the south of England, for example, contain courses of septarian nodules (see SEPTARIA), which have been in great request for making the best kinds of Roman cement. They are concretions of impure calcareous matter, many of them having this analysis: Carbonate of lime, 66; silica, 18; alumina, 7; and protoxide of iron, 6; or consist of these substances in nearly that proportion. Cement-stones are carefully calcined in kilns, and afterwards ground and sifted. Good Roman cement should set in about 15 minutes, and this quick-setting property makes it valuable for work which requires to be executed between tides and for other purposes where the cement used must harden quickly. It is at best of but medium streth. Some natural