[image ALT: Much of my site will be useless to you if you've got the images turned off!]
Bill Thayer

[image ALT: Cliccare qui per una pagina di aiuto in Italiano.]

[Link to a series of help pages]
[Link to the next level up]
[Link to my homepage]

[image ALT: link to previous section]
This webpage reproduces a chapter of
The Cities and Cemeteries of Etruria

by George Dennis

published by John Murray, Albemarle Street
London, 1848.

The text is in the public domain.


[image ALT: link to next section]

Chapter XLVIII



[image ALT: missingALT]
1 Ancient gates
2 Probable site of a gate
3, 3 Square towers, external and internal
4, 4 Circular towers, internal
5 Round tower of Roman work
6 The Acropolis
7 Ruins — Etruscan, Roman, and mediaeval
8 Deep pit, perhaps a quarry
9 Roman columbarium

[image ALT: missingALT]

Cernimus antiquas nullo custode ruinas
Et desolatae moenia foeda Cosae.


Go round about her, and tell the towers thereof.
Mark well her bulwarks; that ye may tell them that come after.


Thayer's Note: Cosa has been conclusively shown not to be Etruscan at all, but rather a Roman foundation of the 3c B.C.

As Cosa was in the time of the Emperor Honorius, such is it still — a deserted waste of ruins, inclosed by dilapidated walls; fourteen centuries have wrought no change in its condition. Yet it is one of the most remarkable Etruscan sites, and should not fail to be visited by every one interested in ancient fortifications.

It occupies the flat summit of a truncated conical hill,  p270 about six hundred feet high, which from its isolation, and proximity to the sea, forms a conspicuous object in the scenery of this coast. It stands just outside the Feniglia, the southernmost of the two necks of sand which unite Monte Argentaro to the south-east of Orbetello.​1 It were best to leave the high-road, where it begins to rise at the foot of the hill of Cosa, and turn down a lane to the right. You will presently perceive a lonely house in a garden, called La Selciatella, the only habitation hereabouts. Here you can leave your vehicle; but if you have a cavalcatura you need not dismount — only ask for one Pietro Fruggioni, who dwells here, and will act as your guide to the ruins; and a more obliging, civil-spoken cicerone you will nowhere meet. Some travellers who have visited Cosa have followed the high road to the further side of the city, and taken as their guide a soldier from the Torre della Tagliata; but this is unnecessary, for Pietro knows the site as well as any one, having tended his cattle there for many a year, and can point out all the lions, which as much as can be expected from these country ciceroni; the traveller must exercise his own judgment as to their origin, antiquity, and purpose. Enquire not for "Cosa," or you will be answered by a stare of surprise, and "non c'è quì tal roba," but for "Ansedonia," the modern appellation of the site.

It is a steep ascent of a mile or more to the walls of  p271 Cosa. You may trace the ancient road all the way to the gate, running in a straight line up the rocky slope; it is but a skeleton, marked by the kerb-stones, for the inner blocks are in few places remaining. On the way it passes some Roman ruins of brick, among them a columbarium.

He who has not seen the so‑called Cyclopean cities of Latium and Sabina, of Greece and of Asia Minor, those marvels of early art, which overpower the mind with their grandeur, bewilder it with amazement, or excite it to active speculations as to their antiquity, the race which erected them, and the state of society which demanded fortifications so stupendous on sites so inaccessible as they in general occupy; — he who has not behed those sublime trophies of early Italian civilization — the bastion and round tower of Norba — the gates of Segni and Arpino — the citadel of Alatri — the many terraces of Cora — the covered way of Praeneste, and the colossal works of the same masonry in the mountains of Latium, Sabina, and Samnium, will be astonished at the first view of the walls of Cosa. Nay, he who is no stranger to this style of masonry, will be surprised to see it on this spot, so remote from the district which seems its peculiar locality. He will behold in these walls immense blocks of stone, iregular polygons in form, not bound together with cement, yet fitted with so admirable nicety, that the joints are mere lines, into which he might often in vain attempt to insert a penknife: the surface smooth as a billiard-table; and the whole resembling, at a little distance, a freshly plastered wall, scratched over with strange diagrams.

The form of the ancient city is a rude quadrangle, scarcely a mile in circuit.​2 The walls vary from twelve  p272 to thirty feet in height, and are relieved, at intervals, by square towers, projecting from eleven to fifteen feet, and of more horizontal masonry than the rest of the fortifications. Fourteen of these towers, square and external, and two internal and circular, are now standing, or to be traced;​3 but there were probably more, for in several places are immense heaps of ruins, though whether of towers, or of the wall itself fallen outwards, it is difficult to determine.

Though Cosa resembles many other ancient sites in Italy in the character of its masonry, it has certain peculiarities. I remember no other instances of towers in polygonal fortifications, with the exceptions of the bastion and round tower of Norba, a similar bastion at Alatri, near the Porta S. Francesco, and the towers at Fondi, apparently of no high antiquity.​4 In no case is there a continuous chain of towers, as round the southern and western walls of Cosa. Another peculiarity of these fortifications  p273 is, that in many parts they rise above the level of the area they enclose, as is also the case at Volterra and Rusellae; whereas the walls of the Latin and Sabine towns are generally mere embankments.​5 The outer half of the wall also is raised three or four feet above the inner, to serve as a rampart: this I have seen on no other site. The total thickness of the wall in this superficial part is between five and six feet. The inner surface is not smoothed like the outer, but left in its natural state, untouched by hammer or chisel; showing in the same piece of walling the rudest and the most finished styles of Cyclopean masonry, and bearing testimony that the outer surface was hewn to its perfection of smoothness after the blocks were raised. A fourth peculiarity is, that while the lower portions of the walls are of decidedly polygonal masonry, the upper parts are often composed of horizontal courses, with a strong tendency to rectangularity, and the blocks are generally of smaller dimensions than the polygonal masses below them. The line between these different styles is sometimes very decidedly marked, which seems confirmatory of the notion suggested by the first sight of this masonry, that it is of two different epochs; the rectangular marking the repairs — a notion further strengthened by the fact, that the material is the same throughout — a close grey limestone. For if the peculiar cleavage of the rock had led to the adoption of the polygonal style in the first instance, it would continue to do so throughout; and any deviation from that style would seem to have been the work of another race, or subsequent age. On the  p274 other hand it may be said, that this rectangular masonry is but the natural finishing off of the polygonal, just as the latter generally runs into the horizontal at angles, as may be observed in the gates and towers of this same city.6

From the ramparts you may perceive that the walls fall back in some degree, though never so much as in a modern revêtement, but the towers are perpendicular on every side, save in a few cases where the masonry is dislocated, and they topple over.7

Of gates there is the orthodox number of three; one in the centre of the northern, southern, and eastern walls of the city respectively.​8 They are well worthy of attention, all of them being double, like the two celebrated gateways of Volterra, though without even the vestige of an arch. The most perfect is that in the eastern wall, which is represented in the woodcut at the head of this chapter.9  p275 It is evident that it was never arched, for the door-post still standing rises to the height of nearly twenty feet in a perfectly upright surface; and as in the Porta di Diana of Volterra, it seems to have been spanned by a lintel of wood, for at the height of twelve or fifteen feet is a square hole, as if for its insertion.​10 The arch indeed is never found, in Italy at least, in connection with this style of masonry; but the gateways of Cyclopean cities were either spanned by flat slabs of stone, or when of too great a width, by lintels of wood, or else by stones overlapping each other, and gradually converging till they met and formed a rude sort of Gothic arch.11

The other two gateways, though more dilapidated, show that they have been formed on the same plan as this in the eastern wall. In the one to the south is a block, nine feet by four, the largest I observed in the walls of Cosa. In this gate also is a large round hole in the inner doorpost for the insertion of a wooden lintel.

I observed no instances of sewers opening in these walls, as usual in Etruscan fortifications, and as are found also in  p276 certain other Cyclopean cities of Italy.​12 Yet such may exist, for I found it impossible fully to inspect the walls on the southern and western sides, the slopes beneath them being covered with a wood so dense as to be often impenetrable, thought the difficulties are not aggravated, as at Rusellae, by any thickets more formidable than myrtle, lentiscus, and laurestinus.

Within the city, all is ruin — a chaos of crumbling walls, overturned masonry, scattered masses of bare rock, and subterranean vaults, "where the owl peeps deeming it midnight," — all overrun with shrubs and creepers, and acanthus in great profusion. The popular superstition may be pardoned for regarding this as the haunt of demons; for ages it was the den of bandits and outlaws, and tradition, kept alive by the natural gloominess of the spot, has thus preserved, it may be, the remembrance of their atrocities. At the south-western corner of the area was the Arx, for the ground here rises considerably above the ordinary level, and is banked up with masonry in parts polygonal, but in general regular, like that in similar situations at Rusellae. On this platform are several ruins, bare walls rising to the height of twenty feet, apparently of the low Empire, or still later, of the middle ages; and numerous foundations, some of the same small cemented masonry, others of larger rectangular blocks, decidedly Roman, and some even polygonal, like the city-walls. It is probable  p277 that the latter, as the earliest masonry — for in many parts the Roman work rests on it — marks the foundations of the three temples which the Etruscans were wont to raise in every city to the divine trio, Jupiter, Juno, and Minerva.13

Within the gate to the east, are many remains of buildings, some with upper stories and windows; and not far from this is a deep hollow with precipitous walls of rock, which seems to have been a quarry.

Joyfully will the traveller hail the view from the ramparts of Cosa; and in truth it were hard to find one on this coast more singular, varied, and grand. Inland, rise lofty walls of rock — rugged, stern, and forbidding — blocking up all view in that direction. At his feet spreads the sun-bright bay, with Porto Ercole and its rocky islet on the further shore,​14 but not a skiff to break the blue calm of its waters; the wide lagoon is mapped out by its side; and the vast double-peaked mass of Monte Argentaro, the natural Gibraltar of Tuscany, overshadows all, lying like a majestic vessel along the shore, moored by its three ropes of sand​15 — the castellated Orbetello being but a knot in the centre of the middle one. To the north he looks along the pine-fringed coast to the twin headlands of the bay of Telamone, and then far away over the level Maremma, to the distant heights of Troja and the  p278 grey peaks of Elba. The Giglio, the so called "Lily" island, is lost behind the Argentaro; but, as it travels southwards, the eye rests on the islet of the Giannutri;​16 and, after scanning the wide horizon of waters, meets land again in the dim hills above Civita Vecchia. The intervening tract is low, flat, desert, — here a broad strip of sand, — there a long, sea-shore lagoon, or a deadly fen or swamp, — now a tract dark with underwood, — now a wide, barren moor, treeless, houseless —

Arsiccia, nuda, sterile, e deserta.

Yet in this region, all desolate as it now appears, stood Vulci, that mine of sepulchral treasures, and Tarquinii, the queen of Etruscan cities, with her port of Graviscae; and Corneto, her modern representative, may be descried, thirty miles off, lifting her diadem of towers above the nearer turrets of Montalto.

Around the walls of Cosa there are few relics of antiquity. It is said that in the plain below are "very extensive remains of a wall of much ruder construction" than those of the city;​17 but I did not perceive them. Near the Torre della Tagliata are several ruins of Roman date, of which those commonly called Bagni della Regina are the most remarkable. You enter a long cleft in the rock, sixty or seventy feet deep, and on one side perceive a huge cave, within which is a second, still larger, apparently formed for baths; for there are seats cut out of the living rock — vivo sedilia saxo — but all now in utter ruin. The place, it has been remarked, recalls the grotto of the Nymphs, described by Virgil;​18 but popular tradition has peopled it with demons, as says Faccio degli Uberti —

Ivi è ancor ove fue la Sendonia,
Ivi è la cava, ove andarno a torme,
Si crede il tristo, overo le demonia.

 p279  Among the ruins on the shore at this spot is some mosaic pavement. The site has been taken, with considerable probability, for that of Subcosa.19

No tombs are to be seen on the slopes around Cosa.​20 It is probable, that, like the one at Rusellae, and those of Cortona and Saturnia, they were constructed of rude masonry, and covered over with earth. Such seems to have been the plan adopted on sites where the rock was too hard to admit of easy excavation. At Volterra and Populonia it was not necessary, for there were soft strata in the neighbourhood.

The walls of Cosa, so unlike those of most cities of Etruria, to what people, and to what age shall we refer them? Can it be that they were raised by the Etruscans themselves — induced to depart from their general style of masonry by the local rock having a natural cleavage into polygons? Or are the peculiarities of these and similar walls in Etruria characteristic of the race which constructed them, rather than of the materials of which they are formed? Are they to be attributed to the earliest occupants of the land, the Umbri or the Pelasgi? — or to much later times, and to the Roman conquerors? The latter view seems now in favour. It was first broached by Micali, the great advocate of the indigenous origin of the Etruscans, and who sought, by invalidating the antiquity of this polygonal style, to enhance that of the regular masonry, which is more peculiarly Etruscan. He  p280 maintains that the walls of Cosa, and of Saturnia, which resemble them, are among the least ancient in the land; and he suggests that they may have been raised by the Roman colony, established here at the close of the fifth century of the City, seeing that the Romans are known to have employed this masonry in certain of their public works.21

It would demand more room than the limits of this work will allow, to discuss this subject to its full extent. But I must make a few remarks.

This polygonal masonry is of high antiquity, long prior to Roman times, though every instance of it cannot claim to be of so remote a date. It must, however, be of later origin than that composed of unhewn masses, rudely piled up, with no further adjustment than the insertion of small blocks in the interstices — that style which, from the description of Pausanias, is sometimes designated "Cyclopean;"​22 for this polygonal masonry is the perfecting  p281 of that ruder mode of construction.​23 Yet that this smooth-surfaced, closely-joined style, as seen in the walls of Cosa, is also of early origin, is proved, not only by the numerous instances of it on very ancient sites in Greece and Italy — some referred to as marvels of antiquity by the ancients themselves — but also by the primitive style of its gateways, and the absence of the arch in connection with it.​24 The fact of the Romans adopting this style of masonry, as they seem to have done in the substructions of some of their great Ways, and perhaps in a few cities of Latium,​25 in no way militates against the high antiquity of the type. The Romans of early times were a servile race of imitators, who had little original beyond their  p282 bellipotentia, and were ever borrowing of their neighbours, not only civil and religious institutions, and whatever ministered to luxury and enjoyment, but even the sterner arts of war. Thus in their architecture and fortifications: in Sabina they seem to have copied the style of the Sabines, in Latium of the Latins, in Etruria of the Etruscans. How much they may have been led to this by the local materials, is a question for separate consideration.

Conceding that the style of masonry must to a considerable extent have been affected by the character of the materials employed, I cannot hold, with some, that it was the natural and unavoidable result — I cannot believe in a constructive necessity — that with certain given materials every people in every age would have produced the same or a similar description of masonry. There are conventionalities and fashions in this as in other arts. It were easy, indeed, to admit the proposition in regard to the ruder Cyclopean style, which is a mere random piling of masses as detached from the quarry; a style which may suggest itself to any people, and which is adopted, though on a much smaller scale, in the formation of fences or of embankments by the modern Italians and Tyrolese, and even by the peasantry of England and Scotland, on spots where stone is cheaper than wood. But the polygonal masonry of which we are treating stands on a totally different ground; and it seems unreasonable to suppose that the marvellous neatness, the artistic perfection displayed in polygonal structures like the walls of Cosa, could have been produced by any people indifferently who happened to fix on the site. For it is not the mere cleavage of the rock into polygonal masses that will produce this masonry. There is also the accurate and laborious adjustment, the careful adaptation of parts, and the subsequent smoothing of the whole into an uniform, level surface. If  p283 ever masonry had the stamp of peculiarity it is this. Not the regular isodomon of the Greeks, nor the opus reticulatum of the Romans has it more strongly marked. I could as readily believe that the Corinthian capital was invented by every nation by which it has been adopted, as that this style of masonry had an independent origin in every country where it has been found.26

The question next arises, to what particular race is this peculiar masonry to be ascribed. No doubt when once introduced, the fashion might be adopted by other tribes than that which originated it,​27 but the type, whose source alone we are considering, would still be proper to one race. Now at the risk of being thought to entertain old-fashioned opinions, I must confess that I can refer it to no other than  p284 the Pelasgi. Not that, with Sir W. Gell, I would cite the myth of Lycaon, son of Pelasgus, and founder of Lycosura, as proof that this masonry was of Pelasgic origin​28 — I might even admit that "there is no conclusive evidence in any one instance of the Pelasgian origin of the monuments under consideration,"​29 — yet the wide-spread existence of remains of this masonry through the countries of the ancient world, the equally wide diffusion of the Pelasgic race,​30 and the remarkable correspondence of the lands it occupied or inhabited with those where these monuments most abound; to say nothing of the impossibility of ascribing them with a shadow of reason to any other particular people mentioned in history — afford satisfactory evidence to my mind of the Pelasgic origin of the polygonal masonry. And here it is not necessary to determine the much vexata quaestio, what and whence was that Pelasgic race, which was so widely diffused throughout the ancient world; it is enough to know that in almost every land which it is said to have occupied, we find remains of this description.​31 In Thessaly, Epirus, and the Peloponnesus,  p285 the peculiar homes of this people, such monuments are most abundant; they are found also in the Isles of the Aegean Sea, and on the coast of Asia Minor, which were at some period occupied or colonised by the Pelasgi. In Italy also, those regions which abound most in such monuments were all once in possession of the Pelasgi, though it must be acknowledged on the other hand, that we have historic mention of that race in certain other districts — at the head of the Adriatic, and in Oenotria — where no such remains have been discovered;​32 nor indeed do we find walls of this character in all the ancient cities of central Italy — even of Etruria — which are said to have had a Pelasgic origin.​33 These discrepancies, whether real or apparent, whether occasioned by the character of the local rock,​34 or by the entire destruction of the earliest  p286 monuments of the land, are but exceptions to the rule, and do not invalidate the evidence for the Pelasgic origin of this peculiar masonry.

With respect to Cosa, there is no reason whatever for regarding its walls as of Roman construction. There is nothing which marks them as more recent than any other ancient fortifications in Italy of similar masonry. The resemblance of the gateways to those of Volterra, and the absence of the arch, point to a much earlier date than the establishment of the Roman colony, only two hundred and seventy-three years before Christ; but whether they were erected by the Pelasgi, or by the Etruscans copying the masonry of their predecessors, is open to doubt. As the walls of Pyrgi and Saturnia, known Pelasgic sites, were of the same polygonal construction; it is no unfair inference that these of Cosa, which has relation to the one by proximity, to the other by situation on the coast, are of a like origin. The high antiquity of Cosa is indeed attested  p287 by Virgil, when he represents it, with other very ancient towns of Etruria, sending assistance to Aeneas.​35 Some, however, have inferred from Pliny's expression — Cossa Volcientium — that it was a mere colony of Vulci, and one of the latest of Etruscan cities;​36 but Niebuhr with more probability considered that the original inhabitants of Cosa were not Etruscans, but an earlier race who had maintained their ground against that people.​37 The connection indeed between Vulci or Volci, and Volsci, is obvious, and from the fact that at one time the Etruscans  p288 possessed the land of the Volsci, it would seem that this was not one of name merely,​38 But the Volsci were of Opican or Oscan race, and what affinity existed between them and the Pelasgi is doubtful; whether an affinity of origin, or one arising merely from the occupation of the same territory at different epochs. Confusion of names and races on such grounds is common enough in the records of early Italy. As the Etruscans were frequently confounded with their predecessors, the Tyrrhenes, so the Volsci may have been with the Pelasgi.​39 It is well known that walls precisely similar to these of Cosa abound in the territory of the Volsci, but whether erected by the Pelasgi, by the Volsci themselves, or by their Roman conquerors, is still matter of dispute; yet by none are they assigned to a later date than the reign of Tarquinius Superbus, two centuries and a half before the Roman colonization of Cosa, which was in the year 481.​40 I repeat that there is no solid ground whatever for ascribing these polygonal walls of Cosa to so recent  p289 a period. With just as much propriety might the massive fortifications of Paestum, which was colonised in the same year, be referred to the Romans.41

Beyond the mention made by Virgil, which can only be received as evidence of her high antiquity, we have no record of Cosa in the days of Etruscan independence. She probably fell under the Roman yoke at the same time as Vulci — on or soon after the year 474 (B.C. 280).​42 Her fidelity during the Second Punic War, when with seventeen other colonies she came forward and saved the Republic, at a time when Sutrium, Nepete, and other colonies refused their aid, is highly commended by Livy.​43 At what period the city was deserted, and fell into the utter ruin which was witnessed by Rutilius at the commencement of the fifth century after Christ, we knowº not;​44 we only learn from the same poet the traditional  p290 cause of such desolation, with needless apologies for its absurdity. The mountain laboured and brought forth, not one "ridiculous mouse," but so many as to drive the citizens from their fire-sides —

Ridiculam cladis pudet inter seria causam

Promere, sed risum dissimulare piget.
Dicuntur cives quondam migrare coacti

Muribus infestos deseruisse lares.
Credere maluerim pygmeae damna cohortis,

Et conjuratas in sua bella grues.

The Author's Notes:

1 The site of Cosa has been much disputed. Some have placed it at Orbetello, others at Santa Liberata, near Santo Stefano on Monte Argentaro; yet Strabo (V. p225) has described its position so as to leave no reasonable doubt of its whereabouts. "Cossa, a city a little above the sea. The lofty height on which the town is situated lies in a bay. Below, lies the Portus Herculis, and hard by, the sea-marsh; and on the headland which overhangs the bay is a tower for watching the tunny-fish." He also states that Cossa is 300 stadia (37½ miles) from Graviscae; and from Populonium nearly 800 stadia (100 miles), though some say 600 stadia (75 miles). Cf. Rutil. Itin. I.285 et seq.

2 Micali's Plan of the city, from which that annexed is adapted, makes it about 2,640 braccia, or 5,060 feet English, in circumference.

3 On the northern side there is but one tower and that in a ruined state; but on the western, or that facing the sea, which was most open to attack, I counted, besides a circular one within the walls, seven external, in various states of preservation, the southmost being the largest and most perfect. This tower is 22 feet wide, and about 20 feet high, as it now stands. In the wall to the south are five towers square and external, and one, internal and circular, 42 feet in diameter. On the eastern side there is but one ancient square tower, and one semicircular of smaller and more recent masonry. Though I have called these towers external, they also project a little inward, from the line of walls. In Micali's plan many of these towers are omitted.

It will be observed that here, as at Falerii, the external towers are not of that form recommended by Vitruvius (I.5), who says they should be either round or many-sided, for the square ones are easily knocked to pieces by the battering-ram, whereas on the circular it can make no impression. The weakness of square towers, however, was ascertained long before the time of Vitruvius; for in one of the very early and curious Assyrian reliefs from the ruins of Nineveh, recently placed in the British Museum, which represents the siege of a city, the battering-ram is directed against the angles of a tower, from which it is fast dislodging the blocks.

4 Memor. Inst. III. p90. Even Pyrgi, which was fortified with similar masonry, though its name signified "towers," retains no trace of such in its walls (ut supra, page 16).

5 I have visited most of those ancient cities in the mountains of Latium, and in the land of the Aequi, Volsci, and Hernici, and remember no other instance than the round tower at Norba, which rises above the level of the city. The height of the eastern wall of Cosa above that level varies from a few feet to twelve or fifteen, and externally the wall is at least double that height.

6 These features are shown in the woodcut at the head of this Chapter, which represents the eastern gate of Cosa. The masonry, though decidedly polygonal, appears in the door-post of the gate to be rectangular. In the fragment of walling to the left, the blocks are polygonal below, and regular above, or at least laid in horizontal courses. The manner in which small pieces were fitted into the interstices is also shown. But the peculiarities of the masonry are not so striking in this, as in many other portions of the fortifications. It was selected from several sketches, as illustrative also of the gate. On this side of the city the masonry is smaller than on the others. The largest of the blocks in the woodcut is not more than 4 feet square, and the height of the wall is only 15 or 16 feet.

7 The bastion and round tower of Norba, on the contrary, narrow upwards considerably.

8 There may have been a postern in the south-eastern angle of the walls, at the spot marked 2 in the Plan. Sir R. C. Hoare also thought he could perceive four gates; and he speaks of four ancient roads. Classical Tour, I. p58.

9 Its entrance is about 12 feet wide, but the passage within is double that in width and 28 feet long; the inner gate is no longer standing, though indications of it are traceable. The depth of the outer doorposts, or in other words the thickness of the wall, is 7 feet, 8 inches. Gateways on a similar plan are found in the Cyclopean cities of Latium — the Porta di S. Francesco at Alatri, and the Porta Cassamara at Ferentino for instance; the latter however is probably of Roman construction.

The gates of Cosa, unlike those of Volterra, do not exemplify the precepts of Vitruvius (I.5), that the road to a gateway should be so arranged, that the approaching foes should have his right side, or that unprotected by his shield, open to the attacks of the besieged.

10 It is shown in the woodcut, together with the upright groove for the saracinesca, or portcullis, like that in the Porta all'Arco of Volterra.

11 In Greece, however, regularly arched gateways have been found in connection with this polygonal masonry. At Oeniadae, in Acarnania, is a postern of a perfect arch in the polygonal walls of the city. Leake, Northern Greece, III. pp560 et seq.; Mure, Tour in Greece, I. p109; and Ann. Inst. 1838, p134. Mon. Ined. Inst. II. tav. LVII. And at Xerokampo, in the neighbourhood of Sparta, is a bridge on the true arch-principle, in the midst of masonry of irregular polygons, though of unusually small size. It was discovered by Dr. Ross of Athens, but first made known to the world by Colonel Mure, in the Ann. Inst. 1838, p140; Mon. Ined. Inst. loc. cit.; and afterwards in his interesting Tour in Greece, II. p248. Several archaeologists of eminence, however, who have seen it have declared to me their full conviction that this bridge is of late date and of Roman construction. Cf. Bull. Inst. 1843, p77. In the polygonal walls of Oenoanda in the Cibyratis, north of Lycia, there is a gateway regularly arched, with Greek inscriptions on tablets in the masonry by its side; as I learn from the portfolio of Mr. Edward Falkener.

12 Besides the instances of such openings in the walls of Norba, Segni, and Alatri, referred to in a former Chapter (see page 121), I may mention a sewer in the walls of the latter city, close to the bastion by the Porta di San Francesco, which is of very peculiar form — a truncated cone inverted, apparently 2 feet wide above, tapering to 1 foot below, and about 3 feet in height. The better known opening in the walls of the citadel of Alatri, I do not believe to be a sewer, but a postern. In the Cyclopean walls of Verulae, now Veroli, in the rudest and most ancient parts of the masonry, are several sewers — tall upright openings, like that in the walls of Norba, or yet more similar in form and dimensions to those so common in the cities of southern Etruria.

13 Servius, ad Virg. Aen. I.422.

14 The Portus Herculis of Rutilius (I.293), and the Itineraries. It was also called Portus Cosanus. Liv. XXII.11; XXX.39. I did not visit it; but Sir R. C. Hoare says it is a singular town, and "resembles a flight of steps, each street bearing the appearance of a landing-place." Classical Tour, I. p56. There are said to be no antiquities remaining. Viag. Ant. per la Via Aurelia, p54.

15 It is highly probable that the Monte Argentaro was once an island; but it is difficult to account for the formation of the two isthmi. The Tombolo, or that to the north, may have been deposited by the Albegna, which opens hard by; but for the Feniglia — there is no river discharging itself hereabouts. The circuit of 36 miles, which Rutilius (I.318) ascribes to this promontory, is much exaggerated. For the physical features and productions of this singular district, see Brocchi, Osservazioni naturali sul promontorio Argentaro, Bibliot. Ital. XI., and Repetti, s. v. Orbetello.

16 The Dianium, or Artemisia of the ancients. Mela, II.7; Plin. III.12.

17 Classical Museum, V. p180.

18 Aen. I.167; Repetti, III. p679.

19 Mannert, Geog. p366. According to this writer, it is this spot which is called Ansedonia, and not the ruined city above. Holstenius (Annot. ad Cluver. p30) made the same distinction; but both seem to have been led to this conclusion by the lines of Faccio degli Uberti, quoted above; for the city itself is certainly now called Ansedonia.

20 Yet excavations have been made in the neighbourhood. Micali (Mon. Ined. p328) states that what was found here in 1837, was presented by himself to the late Pope; and speaks of a flat vessel of bronze, containing an odoriferous gum, which, when burnt, gave forth a most agreeable perfume.

21 Micali, Ant. Pop. Ital. II pp144, 196; III. p6. "A mere glance," he says, "at the walls of Cosa, so smooth and well preserved, proves their construction to be of small antiquity in comparison with those of Fiesole and Volterra, of quadrilateral blocks, and of genuine Etruscan workman­ship." The superior sharpness and freshness in these walls of Cosa, however, are no proof whatever of a less remote antiquity. Micali's argument, to have any weight, should show that the material of which these walls are respectively composed, is either the same, or one equally affected by atmospheric influences. Whereas the fortifications of Volterra and Fiesole, and, it may be added, of Populonia and Cortona, are either of macigno, stratified sandstone, or of other rock equally friable, while those of Cosa and Saturnia are respectively of hard limestone and travertine. I cite Micali in this instance, not as the writer who has treated the subject in the most able manner, but as the originator of the opinion of the Roman origin of Cosa, and as one who has been referred to as authority on the point.

22 Pausan. II16, 4; 25, 7; VII.25. Pausanias, however, applies the same term to the walls of Mycenae, which are of hewn polygonal blocks, and even to the celebrated Gate of the Lions, which is of regular, squared masonry. The term is also repeatedly used by Euripides, in reference to the walls of Mycenae, or of Argos (Elect. 1158; Iphig. Aul. 152, 534, 1501; Orest. 963; Troad. 1083; Herc. Fur. 944; compare Seneca, Herc. Fur. 997; Statius, Theb. I.252). It is therefore clear that the term (p281)"Cyclopean" cannot with propriety be confined, as it has been by Dodwell, Gell, and others, to masonry of the rudest unhewn description, in contradistinction to the neater polygonal, or to the horizontal style. The term was employed in reference to the traditions of the Greeks, rather than to the character of the masonry; or if used in this way it was generic, not specific; applicable to any walling of great massiveness, which had the appearance, or the reputation of high antiquity. "Arces Cyclopum autem, aut quas Cyclopes fecerunt, aut magni ac miri operis; nam quicquid magnitudine suâ nobile est Cyclopum manu dicitur fabricatum." Lactant. ad Stat. Theb. I.252; cf. I.630. Though rejected altogether by Bunsen (Ann. Inst. 1834, p145), the term is convenient — se non è vero, è ben trovato — and in default of a better, has some claim to be retained. On this ground I have made use of it in the course of this work in its generic sense, applying it alike to all early massive irregular masonry.

23 Gell held the contrary opinion — that the polygonal was more ancient by some centuries. Topog. Rome, II. p165.

24 Gerhard (Ann. Inst. 1829, p40), remarking on this fact, says it seems certain that even the least ancient remains of this description preceded the invention of the arch. But this is refuted by the recent discovery of arches in connection with this masonry in Greece and Asia Minor. Ut supra, page 275. In none of these cases, however, have the structures an appearance of very remote antiquity.

25 In the Via Salaria, near Rieti, and in several places between Antrodoco and Civita Ducale; in the Via Valeria, below Roviano, and elsewhere between Tivoli and Tagliacozzo; and in the Via Appia, between Terracina and Fondi. The cities, whose polygonal fortifications have been ascribed to the Romans, are Norba and Signia. Gerhard, Ann. Inst. 1829, p55, et seq.; 83, et seq.; Bunsen, Ann. Inst. 1834, p144; Bunbury, Classical Museum, V. p167, et seq. Strabo (V. p237) states that most of the cities on the Via Latina, in the lands of the Hernici, Aequi, and Volsci, were built by the Romans.

Thayer's Note: The suspicion that the polygonal walls of Norba were in fact Roman (and thus not of high antiquity) seems to have been confirmed by excavation in the early 20c: see Pfeiffer & Ashby, Suppl. Papers, Am. Sch. of Class. Studies, Vol. I, pp89‑90.

26 The adoption of this style by the Romans in the pavements of their high-ways, in no way affects the question. The earliest of these roads, the Via Appia, was constructed only in the year 442 (B.C. 312) — ages later even than those polygonal cities which are sometimes ascribed to the Romans; and it may be that they but imitated the roads of their predecessors. Still less can the use of polygonal pavement by the modern Florentines, be admitted as an argument against the peculiarity of the type, as Micali would fain have it. Ant. Pop. Ital. I p197. They have but adhered to the style which was handed down to them from antiquity, while the modern Romans have preferred the opus reticulatum, as the model for their pavements. And though Micali contends for a constructive necessity, it is completely set aside by the fact, which he mentions, that the stone for the pavement of Florence is brought from the heights of Fiesole; for the horizontal cleavage of that rock is most manifest and notorious.

Nor can the existence of polygonal masonry in the fortresses and other structures of the aboriginal Peruvians, be regarded as opposed to the peculiarity of the type. Too great a mystery hangs over the origin of that singular race, and of its civilization, for us to admit them as evidence in this question. The style seems to have differed from that of the polygonal masonry of the old world, resembling it in little more than the close-fitting of the masses. If anything is to be learned from these structures, it is that they contradict the doctrine of a constructive necessity; being of granite or porphyry, which have no polygonal cleavage; and are rather suggestive of a traditional custom. See Prescott's Conquest of Peru, I. pp16, 143.

27 Chevalier Bunsen maintains that many of the polygonal fortifications of Italy were raised by the Volsci, Aequi, and Hernici. Ann. Inst., 1834, p142. But if this be admitted, it does not prove that the type originated with them.

28 Gell, Rome, II. v. Pelasgi.

29 Bunbury, Clas. Mus. V. p186. Yet there is, in most instances, the same kind and degree of evidence as lead us to ascribe the walls of Fiesole and Volterra to the Etruscans, those of Paestum to the Greeks, or Stonehenge to the Druids. We find it recorded that in very early times the lands or sites were occupied by certain races; and finding local remains, which analogy marks as of high antiquity, and not of Roman construction, we feel authorised in ascribing them to the respective people.

Thayer's Note: This was by no means a closed question in the 19c; for a contemporary and opposite opinion, with arguments to back it up, see pp33‑34 of "Avebury, Stonehenge, and Silbury Hill" in Abbeys, Castles and Ancient Halls of England and Wales, Their Legendary Lore and Popular History, by John Timbs (1870, as revised by Alexander Gunn † 1901).

30 "It is not a mere hypothesis," says Niebuhr, "but with a full historical conviction, that I assert, there was a time when the Pelasgians, then perhaps more widely spread than any other people in Europe, extended from the Po and the Arno almost to the Bosphorus." I. p52, Eng. trans.

31 Gerhard (Memor. Inst. III. p72) takes these structures of irregular polygons to be Pelasgic. Müller (Archäologie der Kunst, p27) thinks that most of the so‑called Cyclopean walls of Epirus and the Peloponnesus were erected by the Pelasgi. We know that they built the ancient wall round the Acropolis of Athens; and the way in which this fact is mentioned by Dionysius (I. p22), in connection with their wandering habits, favours the opinion of some, that these Pelasgi were the great fort-builders of antiquity, a migratory race of warlike masons, who went (p285)about from land to land, sword in one hand, hammer and chisel in the other, fortifying themselves wherever they conquered.

32 It is asserted that no polygonal structures are to be found in Basilicata or Calabria; nor, indeed, north of the Ombrone, nor south of the Vulturnus — some say the Silarus. Memor. Inst. I. p72; Ann. Inst., 1834, p143. But, as regards the south of Italy, the assertion is premature. Have sufficient researches been made among the Calabrian Apennines? Petit-Radel, who maintains the Pelasgic construction of this masonry, sserts that there are remains of it far south, in Apulia and Lucania. Memor. Instit. III. pp55‑66. I have heard also, on good authority, that a German gentleman has recently made some singular discoveries of very extensive polygonal remains in this part of Italy, and is about to give an account of them to the world. That no such walls are to be found on the ancient sites at the head of the Adriatic, where the Pelasgi first landed in Italy, may be explained by the nature of the low and swampy coast, which did not furnish the necessary materials.

33 At Falerii, Agylla, and Cortona, which were Pelasgic, we find regular, parallelopiped masonry; at Pyrgi and Saturnia, on the contrary, whose Pelasgic origin is equally well attested, we have remains of purely polygonal construction.

34 It is very probable that the local rock sometimes, though not always, determined the style of the masonry. Where it naturally split into rectangular forms, as is the case with the macigno of Cortona, and the volcanic tufo of southern Etruria, there the horizontal may have been preferred, even by those who were wont to employ a different description of masonry. This seems to have been the case at Agylla, where the rock is of tufo; there are no traces of polygonal construction; even in the most ancient tombs the masonry is rectangular. See page 29. Yet, in spite of these natural (p286)inducements to the contrary, the favourite style was sometimes carried out, as is proved by the tholus of polygonal construction at Volterra, formed of travertine (ut supra, page 160); by the polygonal walls of Saturnia of the same material — a stone of decidedly horizontal cleavage, and used abundantly in regular masonry in all ages, from the Etruscan walls of Clusium and Persia, and the Greek temples of Paestum, to the Colosseum, St. Peter's, and the palaces of modern Rome. This is also proved by the travertine and crag used in the polygonal walls of Pyrgi (see page 12), and by the crag in the similar fortifications of Orbetello (see page 264); and even these walls of Cosa afford abundant proof that the builders were not the slaves of their materials, but exerted a free choice in the adoption of style; for the same stone which was hewn into horizontal masonry in the towers, gateways, and upper courses, as shown in the woodcut at page 269, could have been thrown into the same forms throughout, had not the builders been influenced by some other motive than the natural cleavage. Another singular instance of disregard of cleavage is exhibited in the walls of Empulum, now Ampiglione, near Tivoli, where the masonry, though of tufo, is decidedly polygonal; this is the only instance known of that volcanic rock being thrown into any other than the rectangular forms it naturally assumes. See Gell's Rome, v. Empulum. These facts will suffice to overthrow the doctrine of a constructive necessity, often applied to this polygonal masonry.

35 Virg. Aen. X.168; Serv. in loc. Müller (Etrusk. I.3, 1) remarks that the walls of Cosa are by no means to be regarded as not Etruscan, because they are polygonal, and considers them as evidence of its antiquity (II.1, 2). Orioli (ap. Inghir. Mon. Etrus. IV.p161) also thinks the walls of Cosa confirm the antiquity assigned to it by Virgil. Abeken (Mittelital. p21) takes Cosa to be Pelasgic; and Gerhard inclines to the same opinion (Ann. Inst. 1831, p205), and reminds us that there was a city of the same name in Thrace. He thinks the name may have and affinity to the Doric κόττα, κοδδά, a head. It is written Cossae by Strabo and Ptolemy, but Cluver (II p479) thinks this was merely owing to the habit of the Greeks of doubling the s in the middle of a word. It is not written so by any Roman author but Pliny, though Virgil gives it a plural termination. If the Etruscan name were analogous it must have been spelt with an uCusa. We find in Etruscan inscriptions the proper names of "Cusis" or "Cusim," "Cusinei," "Cusithia," — Lanzi, II. pp371, 402, 416; Vermigl. Iscriz. Perug. I. p324. "Cusiach" also at Cervetri, (ut supra, page 27), and "Cusu" at Cortona. See Chap. LVI.

36 Plin. III.8. Cluver (II p515), Lanzi (II. p56), Micali (Ant. Pop. Ital. I p147), and Cramer (I. p195), interpret Pliny as saying that Cosa was a colony of Vulci. But the expression he uses is shown by Gerhard to have indicated merely the territory in which a town stood, without reference to its origin; as "Alba Marsorum" signified the Latin colony of Alba in the land of the Marsi. Ann. Inst., 1829, p200. Mr. Bunbury (Classical Museum, V. p180) argues that as Vulci itself did not begin to flourish till after the decline of Tarquinii, for which he cites Gerhard's authority (Ann. Inst., 1831, p101), Cosa, its colony or offset, must needs belong to a late period. But — the question of the colony apart — that Vulci was of so recent a date is wholly unsupported by historic evidence, nay, is refuted by the very archaic character of much of the furniture of its sepulchres. And Müller (Etrusk. II.1, 2) justly observes that Pliny's mention of Cosa does not prove that before it was colonised by the Romans the town had no existence.

37 Niebuhr, I. p120; cf. p70. He founds this opinion on the mention by Livy (XXVII.15) of a people called Volcentes, in connection with the Hirpini and Lucani, whom he took to be of the same race as the Volsci.

38 Cato, ap. Serv. ad Aen. XI.567. The connection between the Etruscans and the Cistiberine people, especially the Oscan races, is very apparent from the names of places. Velathri (Volterra) has its counterpart in Velitrae (Velletri) — Fregenae in FregellaePerusia in FrusinumSutrium in Satricum. A Ferentinum and an Artena existed in both lands; so also a river Clanis. There was a Compsa in Samnium, and a Cossa in Lucania, as well as a river Cosa in the land of the Hernici; and Cora also seems connected with Cosa, the s and r being frequently interchangeable. That the Vulturnus on which Capua stood had an Etruscan name needs no proof. Capua itself is analogous to Capena (Vol. I p175); so is Falerii to Falernus, whose last syllable is merely the ancient adjectival termination. Alatrium seems connected with Velathri, by the dropping of the digamma; so also Aesula with Faesulae. Instances of such analogies might be yet further cited.

39 The names, indeed, bear a strong affinity. Niebuhr (I. p72) points out the analogy between the names Volsci and Falisci; the latter people, he thinks, were Aequi, but they are called in history Pelasgi; and the similarity of the words Falisci and Pelasgi is also striking. (Vol. I p140).

40 Vell.º Paterc. I.14; Liv. Epit. XIV; Cicero (in Verr. VI.61) speaks of Cosa as a municipium. Gerhard suggests that she may have been colonised with the remains of the population of Vulci. Ann. Inst. 1831, p404.

41 If the Romans had any hand in the construction of these walls, it must have been in the upper courses alone, which differ so widely from the lower, though the material is the same throughout. It is possible they may have thus repaired the walls. But if Virgil's testimony as to the antiquity of Cosa be admitted — and who can reject it? — the Romans cannot have raised them entirely, or what has become of the prior fortifications? It is hardly credible that at so early a period they could have been rased to the foundations, so as not to leave a vestige.

42 Vol. I p404.

43 Liv. XXVII.9, 10. She is subsequently mentioned in Roman history. Liv. XXXII.2; XXXIII.24; Caesar, Bell. Civ. I.34; Cicero, ad Attic. IX.11.º Tacitus (Annal. II.39) speaks of Cosa as "a promontory of Etruria." The Emperor Vespasian was brought up in its neighbourhood (Sueton. Vespas. c2); at least Cluver (II p479) and Pitiscus consider the Cosa of Etruria is here meant; but Repetti (I. p829) thinks it is the Cosa of the Hirpini.

44 Rutil. I.285, et seq. Inscriptions, however, prove the city to have been in existence in the middle of the third century of our era. Repetti, I. p828; Reines, III.37, cited by Müller, I. p348.

There are certain coins — with the head of Mars on the obverse, and a horse's head bridled, and the legend Cosano or Cosa on the reverse — which have been attributed to Cosa. Lanzi, II. pp24, 58; Mionnet, Med. Ant. I. p97; Suppl. I. p197. Lanzi infers from the type an analogy with Consus, an equestrian name of Neptune, whence the public games of the Consualia were called (Tertul. de Spect. c5), and thinks Cosa to a Roman must have been equivalent to Posidonia to a Greek. Müller (Etrusk. I p340), who does (p290)not ascribe these coins to Cosa, shows that they cannot in any case belong to the times of the Etruscans, because that people had no O in their language. Cramer (I. p195) refers them to Compsa in Samnium; and so also Millingen (Numis. Anc. Italie, p170); but Sestini (Geog. Numis. II. p4) to Cossea, a city of Thrace.

[image ALT: Valid HTML 4.01.]

Page updated: 31 Oct 17