ground, fo confiderable a portion of the fame pavement was laid ope as, together with other openings, which were made in the courfe of that and the following year, enabled one to afcertain its form and dimenfions. It appeared to have been a fquare of forty-eight feet ten inches. The complete defign of this pavement could now alfo be afcertained, which, for fize and richness of ornaments, is, I believe, equalled by few of thofe difcovered in other provinces of the Roman empire, and is undoubtedly fuperior to any thing of the fame kind hitherto found in this country.' P. 2.

The accuracy of Mr. Lyfons, in diftinguishing, by various modes of drawing and colouring, the different parts of the Mofaics which are ftill exifting, or are taken from other fketches, or only restored from conjecture, must be highly fatisfactory to the curious examiner of his reprefentations. In one inftance it appears to us, that his practice is not quite confiftent with his narrative. Defcribing the principal Mofaic, he fays. "The figures of a gryphon, a bear, a leopard, a flag, a tygrefs, a lion, and a lionefs, are now remaining. Thofe of a boar and a dog, which are to be feen in Mr. Brown's drawing, together with that of an elephant, have been fince deftroyed." He fays afterwards, "The whole defign of this pavement is reftored in plate x, where all the ornaments and figures, which could be precifely afcertained, are defcribed by determined outlines; and thofe parts which are reftored from conjecture, are expreffed only by dotted outlines." According to this, we conceive that the boar and dog, which are ascertained by Mr. Brown's drawing, fhould have been marked by a determined outline; and that the elephant, which is under the fame circumftances, fhould have been fimilarly marked; whereas, the loft parts of the boar and dog are, on the plate, restored in doted outlines, and the elephant, which is exprefsly faid to have been deftroyed, is in no way whatever diftinguished from the figures which are ftill entire. On plate ix, alfo, the elephant appears, next to the dog, as if stilk exifting in the perfect mofaic. These apparent inaccuracies we mention, becaufe, if they really are fuch, it must be well worth the author's while to have them corrected. The female figures in the four angles of the tenth plate, are very elegantly sketched and reftored by Mr. Smirke.

The account of the materials and conftru&ion of the principal Mofaic, is curious and interefting.

"The teffera of which this pavement is compofed, are, for the most part, cubes of half an inch; thofe of the outward border are larger, and those near the centre much fmaller. Many are triangular, and of various other fhapes. The whole when entire, could not, therefore, have contained less than a million and a half of them.

"Moft

"Most of the materials of which they are formed, are the produce of this country, except the white, which are of a very hard calcareous ftone, bearing a good polifh, and nearly refembling the Palombino marble of Italy. Nothing could anfwer their purpofe better than this fubftance, and it was employed by the Romans in many of their Mofaic pavements in other parts of Europe. The dark bluish grey, are of a hard argillaceous ftone, found in many parts of the vale of Gloucester, and there called blue lyas. The afh-colour are of a fimilar kind of stone, and frequently found in the fame mafies with the former. The dark brown are of a gritty ftone, found near Bristol, and in the foreft of Dean. The lightet brown nearly refemble a hard calcareous ftone found at Lypiat, about two miles from Woodchefter. The red are of a fine fort of brick.

"In the autumn of 1794, an opportunity occurred, of exploring the ground to a confiderable depth, near the north-welt corner of the pavement, where it had formerly been broken up. The cement on which the pavement was laid, appeared to be about eight inches thick, and compofed of fine gravel, pounded brick, and lime, forming a very hard fubitance, on which the tefferæ were laid in a fine cement, confitting chiefly of lime. The interfices [between them] appeared to be filled with fo hard a cement, that it was more difficult to break it, than even the ftones of which the tefferæ were formed. The next ftratum was three feet thick, and appeared to be composed of a coarfer gravel, with which great quantities of the teffera were mixed; and below this, another of a reddifh fand and clay, mixed with pieces of brick, about a foot in depth, which lay on the natural foil*."` P. 4.

We shall conclude our account of this very creditable publication, by the author's general defcription of the Roman remains which form its fubject; omitting the chief part of his notes, as before, for the fake of room.

"The feveral buildings above described, appear to be the remains of a Roman house, or rather, perhaps, of a villa; they do not feem, notwithstanding their great extent, to have been part of any town or group of houfes.

"The houfes of the Romans, when fituated in the country, were frequently on one floor; and, if they belonged to perfons of confequence, were of very great extent, and enriched with the moft magnificent and fplendid decorations of every kind, though their external elevation was usually quite plain. They confifted of feveral large halls, porticos, and open courts, running through the centre of the building, with fuites of rooms branching out on either fide.

* Here Mr, Lyfons, in a very fatisfactory note, gives his reader an opportunity of comparing this practice with the precepts of Vitruvius, for making the ruderatio, or substrata of pavements. There is alfo a very good note preceding this, and proving great research, on the difficulty of diftinguishing between fera and fecilia, as ufed by Vitruvius.

"The

"The remains of the building at Woodchefter, bear a ftriking refemblance to the plan of the Roman house above alluded to. They cannot be expected to agree with it in every particular, fince the Romans frequently varied the form of their houfes, to adapt them to the climate of the country in which they were built. The two great courts-run through the middle of the building, and have a great number of rooms of various dimenfions branching from them. The great court on the fouth-fide, of which it is probable that the principal gate of entrance, was fituated as above-mentioned, feems to answer to the periftylium of Vitruvius, and was probably furrounded with a colonnade, though only loofe fragments of columns were found, and none of their bafes could be difcovered, fo as to afcertain their fituations. On the eaft and weft fides of this court, are confiderable ranges of buildings; in the eastern wing of which, the remains of the Laconicum*, are fully fufficient to indicate its original ufe. It is probable, that the room contiguous to it, on the eastern fide, was an apodyteriumt, and that the room in which two stone steps are remaining, was a cold bath, as it was a very common practice with the Romans, to use the cold bath immediately after the fadatory. The ufes of the other rooms in these ranges of buildings, I have not been able to ascertain.

It is very probable, that most of the rooms on the weft fide of the great court, were appropriated to the ufe of the fervants, as they do not appear to have had Mofaic pavements or other decorations. The three large rooms on the north fide of the great court feem, from the fragments of ftatues, marbles, and columns found there, to have been very highly decorated: and, from their fize, it is probable, that they were either aci, or exedrat. The fecond court is furrounded on the north, eaft, and weft fides, with galleries; that on the north fide, has an elegant Mofaic pavement, and a fragment of one remains in that on the east fide. Thefe galleries were clearly what the Romans called cryptoporticus, and the area inclofed within them, might have been the atrium. The room of which the great Mofaic pavement remains, was, no doubt, the cavadium tetraftylon of Vitruvius, and must have been extremely magnificent; for there is great reafon to imagine, from the elegance of the floor, that the cieling, and other parts of the

*«The laconicum, or, as it was fometimes called, offa or calida fudatio, was intended entirely for the purpole of fweating." Note on p. 12. We may add, that they were allo called fudatoria. "Quid cum fudatoriis, in quæ ficcus vapor corpora exhaufturus includitur." Seneca, Epift. 51.

+ Undreffing room.

Rev.

"The aci were a large kind of faloons-ufed for the purposes of entertainments."-The exedra were large rooms of various forms, which are fuppofed to have been furnished with feats, and used for converfation and difputations. Vitruv. lib. v. c. 11.

§ "The cavadia, or cava ædium, appear to have been sometimes large halls, and fometimes open courts in the interior part of the houfe, communicating with feveral fuites of rooms."

room,

room, were richly decorated. It is probable, that part of the roof was formed by diagonal vaultings, refting on the four columns.

"The walls remaining on the weft fide of the cavædium and cryptoporticus, are probably the remains of the triclinia hyberna and baths; as most of them have fubterraneous flues, for the purpose of introducing heat; and their fituation correfponds with that which Vitruvius affigns for thofe apartments, as will appear by the plan. Some of these rooms might alfo have been cubicula, or bedchambers. The apartments on the eaft fide of the cryptoporticus, were probably warmed by the Hypocauft, No. 31, pl. vi. Thefe occupy the fituation affigned by Vitruvius, for the triclinia of the fpring and autumn. It is probable, that most of the Roman houfes in Britain, had fubterra neous flues or hypocaufts, as the nature of the climate must have rendered them occafionally useful at all feasons of the year. The extent of that part of the building which lay on the north fide of the cava dium, cannot now be afcertained, as it has been entirely destroyed bythe erection of the church, except those walls under the chancel, mentioned in p. 6.

"From the magnitude of this building, and the richness of its decoration, it does not feem probable, that it belonged to any private individual. It is more likely that it was a public work built for the refidence of the Proprætor, or at least of the governor of this part of the province, and occafionally, perhaps, of the Emperor himfelf; as it is well known, that several of the Roman Emperors vifited Britain, and fome of them continued there a considerable time." P. 16.

The plates in this work, befides the vignettes and other ornaments, are forty in number, exhibiting, on the whole, a wonderful difplay of Roman work; and, as we cannot repeat too ftrongly, a very honourable proof of the acutenefs, diligence, and talents of Mr. Lyfons.

ART. II. The Ancient Hiftory of Ireland, proved from the Sanferit Books of the Bramins of India. Dedicated to the Prefident and Members of the Royal Irish Academy. By the Author of the Vindication of the Hiftory of Ireland (General Vallancey) 8vo. 30 pp. Dublin. O&. 30, 1797.

FROM the learned labours of our countrymen in India,

who have devoted their leifure hours to the study of Sanferit literature, it appears, that the enquiries of the ancient Bramins, were not confined to the hiftory or mythology of their own nation: thofe venerable fages were acquainted with Irān and Chaldea; they have preserved traditions of the Babylonian

Semiramis,

Semiramis, and recorded the pagan fanctity of Mecca*; indeed, they seem to have extended their researches from the mouths of the Ganges to thofe of the Nilet, and to have acquired a knowledge of all the various intermediate countries.

But however curious thefe difcoveries of Braminical research may be, a more interefting proof of its great extensiveness, is prefented in the publication now announced; from which we learn, that in one of the ancient Puranas, or facred volumes of the Hindoos, a particular defcription has been found of the British Islands. As this circumftance ftrongly corroborates fome affertions and conjectures, published several years ago, by the Vindicator of Irish Antiquities, the paffage in queftion was extracted from the original Purana, and immediately communicated to that learned gentleman, who has annexed a neatlyengraved copy of it, in the Sanfcrit character, to the memoir now before us; which, we hope, is but an introduction to a more confiderable work. The ingenious author, ufing the third perfon, thus begins:

In 1786, General Vallancey published his Vindication of the Ancient Hiftory of Ireland, wherein he fhews the great ignorance of former tranflators in geographical names, and proves one colony there named Tuatha-Dadan, were the Dadanites of Chaldea, defcended from Chus, agreeable to Irish history: that thefe people were known in Oriental hiftory, by the name of Cuthi, and in the Irish history, by that of Aite-Cothi, or ancient Cothi." P. 1.

The coincidence of Sanferit with Irish tradition, and the identity of language, in many inftances, induced the General to communicate his remarks to Sir William Jones, then studying the Sanfcrit at Benares.

"The mysterious characters of the Irish Druids, named Ogham, were particularly recommended to his obfervation: that facred character which none but Druids were permitted to write; and none but Druids were permitted to read. In 1789, Sir William delivered his difcourfe to the academy at Calcutta, printed in the Afiatic Researches, wherein he fhews the word Ogham is pure Sanferit, and means the facred or myfterious writings or language, and that it is used in that fignification, in the books of the Sanfcrit. He alfo obferves, that the Sanferit language was older than the Hindu, was the language of Iran, and of pure Chaldaic origin."-" Mr. Hallis, another learned Orientalift, in a letter dated Benares, 1765, fays, that the Bramins ftudied the Chaldaic language, in which their books of phyfic are chiefly

Afiat. Refearches, vol. iv. Many centuries before the Mohummedan æra, the Black Stone, and various idols, were worshipped at Mecca.

+ Ditto, vol. iii.

written :""