STOP Early Journal Content on JSTOR, Free to Anyone in the World This article is one of nearly 500,000 scholarly works digitized and made freely available to everyone in the world byJSTOR. Known as the Early Journal Content, this set of works include research articles, news, letters, and other writings published in more than 200 of the oldest leading academic journals. The works date from the mid-seventeenth to the early twentieth centuries. We encourage people to read and share the Early Journal Content openly and to tell others that this resource exists. People may post this content online or redistribute in any way for non-commercial purposes. Read more about Early Journal Content at http://about.istor.org/participate-istor/individuals/early- journal-content . JSTOR is a digital library of academic journals, books, and primary source objects. JSTOR helps people discover, use, and build upon a wide range of content through a powerful research and teaching platform, and preserves this content for future generations. JSTOR is part of ITHAKA, a not-for-profit organization that also includes Ithaka S+R and Portico. For more information about JSTOR, please contact support@jstor.org. Classical Philology Volume IX April 10 J 4 Number 2 THE FORM OF THE EARLY ETRUSCAN AND ROMAN HOUSEi By Mabgaret C. Waites Most recent writers agree in their opinions as to the antiquity of what Vitruvius calls the "Tuscan atrium." This type, the one most common at Pompeii, Vitruvius says was roofed by placing beams across the width of the room above the ends of the impluvium, laying cross-beams upon these, and connecting the comers of the rectangle thus formed with the corners of the wall by other timbers laid diagonally.^ This is generally regarded as the oldest type of atrium, probably derived from the East by the Etruscans them- selves and early introduced by them into Rome.' Some authorities, • Thia article was written at the American School of Classical Studies in Rome, and owes much to the encouragement and counsel of the director, Professor J. B. Carter, and Assistant Professor A. W. Van Buren. • Vitruv. vi. 3. 1: "Tuscanica sunt in quibus trabes in atrii latitudine traiectae habeant interpensiva et collicias ab angulis parietum ad angulos tignorum inter- currentes, item asseribus stillicidiorum in medium compluvium deiectis." Cf. also Varro i.L. v. 161: "Tuscanicum (cavum aedium) dictum a Tuscis, posteaquam illorum cavum aedium simulare coepenint." » To show the prevailing view, it may be useful to quote a few examples. Stuart Jones in his Companion to Roman History (1912), 159, says: "Vitruvius describes five types of atria and as that which appears to be the earliest in date and is the most usual at Pompeii is called by him a. Tuscanicum, we may assume that it owes its origin to the Etruscan architects." German authorities are more explicit. Cf. F. Marx, "Die Entwicklung d. rSmischen Hauses," N. Jahrb., XII (1909), 548 ff.: "Die Romer haben .... das Atrium von den Etruskern in der Konigszeit ilbernommen Oben in den engen Bergstadten der etruskischen Konigssitze ist das Atrium entstanden und [Classical Philology IX, April, 1914] 113 114 Margaret C. Waites however, among them Frothingham in Roman Cities (1910), 118 ff., and the writers in Daremberg and Saglio, s.m. "cavum aedium," "domus, " and "tectum," consider the atrium testudinatum, that is, the type with no aperture in the roof, as the earliest form, from which developed first the atrium displutdatum, in which the roof sloped away from the central opening and the rainwater was conducted by gutters to the street, and then the atrium tuscanicum. With this latter view I agree, and I hope in the present article to trace the development from the simpler to the more complicated form of house and to establish the existence of a transition-form between the atrium testvdinaium and subsequent varieties. I For the earliest type of Italic house, one must, as has long been recognized, turn to the hut-urns, found all over Etruria and Latium. They may be round, oval, or more rarely quadrilateral; they contain only one room; they sometimes have windows, real or simulated, on the side-walls; and over the door, for the escape of smoke, there is a small roimd opening, often closed with a disk,i evidently the translation into terra-cotta of a sort of wooden shutter. A corre- sponding opening appears at the opposite end of the ridge-pole. The roof is sometimes roimded in exactly the shape of a testudo,* and sometimes rises in a pyramidal shape' which recalls Varro's description of the atrium, or cavum aedium, testudinatum.* Such a tragi deutlioh das Zeichen seiner Entstehung aus beengten Wohnverhaltnissen Nur das Atrium Tuscanicum genannte Haus verdient die Benennung als Atrium." The latest theory derives the Tuscan atrium from the East; cf. Pfuhl in N. Jahrb. (1911), 173»: "Das griechische Peristyl wie das etruskisch-romische Atrium sind alt-kretisch." See also Pernice, Einleiiung in die Altertumgvdasenschaft (Ed. 1912), II, 25: "Die Form selbst [i.e., of the atrium Tuscanicum] soheint von den Etruskern entlehnt zu sein, jedoch ist es schwer zu sagen wie diese Veranderung ent- standen ist. Die Offnung des Daches erscheint, namlioh, als eine so eigenartige radikale Veranderung, dass man sich schwer zu der Annahme einer Entwicklung vom geschlossenen zum geoffneten Dach entschlieast Die Etrusker werden den neuen Haustypus aus ihrer fruheren Heimat mitgebracht haben." ' Montelius, La civilisation primitive en Italie, Ualie Centrale, PI. 175, No. 14; Altmann, Die Italischen Bundbauten (1906), 14. ' Montelius, op. cit.. No. 15. » Daremberg and Saglio, s.v. "tectum," Fig. 6769. * L.L. v. 161: "Cavum aedium dictum qui locus tectus intra parietes relinque- batur patulus, qui esset ad communem omnium usum. In hoc locus si nuUus reliotus erat, sub divo qui esset, dicebatur Testudo ab testudinis similitudine, ut est in praetorio et castris." The Early Etrxtscak and Roman House 115 house contained, of course, nothing but a common Hving-room, and it is at once obvious that in such restricted space no large opening in the roof over the central hearth could have been tolerated.' In fact the smoke-holes were, as I have remarked, uniformly placed at the ends of the ridge-pole. A more developed and often complicated dwelling is imitated in the Etruscan chamber-tombs, and the question immediately arises whether atria with apertures in the roofs are represented here. One may object that such structures as the chamber-tombs, which are in the large majority of cases subterranean, would naturally be devoid of roof-openings. This objection, however, rests upon a misconception. The tombs were not dwellings, they imitated dwell- ings, and upon the closeness of the imitation the happiness of the dead was probably in large measure thought to depend. If, then, the main room of a dwelling-house had an opening in the center of the roof, and if the roof itself was shaped like that of a displuviate or of a Tuscan atrium, we should expect to find, not necessarily actual openings, but more likely simulated openings in the roofs of the chamber-tombs. The principle leading to the adoption of simulated apertures in the roof would be the same as that which is responsible for the false windows on the side-walls of the hut-urns, for the false doors of the Grotta delle Camere finte^ at Corneto, and for the sham armor of the Regulini-Galassi tomb.' In fact there are, as we shall see, innumerable cases where much ingenuity and labor were expended to make the roof of the tomb an exact imitation of the timbered ceiling of a house. Another more cogent objection urges the conservatism of religion, which would make it quite natural that the architecture of the tomb should lag far behind that of the house. To form a con- ception of early dwellings, however, we have almost no guides but those furnished by religious usage, namely, urns and tombs; for in the very few cases where foimdations of actual Etruscan houses 1 One may query whether it would not have been possible to use a sort of movable curtain or wooden shutter to close an opening in the roof, just as the smoke-hole could be closed. If such an arrangement had ever come into general use, however, the name for this central opening would surely never have been impluvium. Yet such is the term in a passage relating to very early religious usage. See p. 125. 2 Dennis, Cities and Cemeteries of Etruria, I, 364. » Ibid., 267 f. 116 Margaret C. Waites are left to us, it is almost impossible to establish from them con- clusions as to the shape of the roof. Religion, too, may at least be reliable as a relative guide. It may help us to discover which of two styles is probably the earlier, and if in the vast majority of tombs and urns we find this earlier type persisting, only occa- sionally dispossessed by a later development, we shall be warranted in assuming a proportionate, though not an equal, slowness of evo- lution in the dwelling-house. Moreover, there is one instance which shows that, at least occa- sionally, tombs were constructed in the same way as contemporary houses. The houses, namely, of Monte Sant' Angelo in the Faliscan territory show rectangular foundations, but, apparently, a type of roof like that of the round hut. At the same time, tombs in this district, apparently contemporary with the houses, have a similar foundation and a similar style of roofing.* An urn from Corneto mentioned in Notizie degli Scavi (1881), 353 (cf. PI. V, 12 and 13), and a tomb of Vetulonia'^ display the same peculiarity. From these facts, we may perhaps infer that roof-construction both in house and in tomb tended to be conservative. II The simplest type of atrium in a rectangular chamber-tomb had, therefore, under the influence of the hut, a rounded tortoise-shell* 1 Monumenti Antichi dei Lincei, IV (1894), 89 ff. on Tomb VIII: "Si entrava in una camera a pianta trapezoidale. .... Degno di molta considerazione 6 etato per noi ci6 che abbiamo notato nell' imposta della volta. Questa non termina come nelle tombe a camera con intagli nel masso che imitino il soffito di una casa quadrata; ma, come apparisce dalla parte del tufo che non 6 franato nelle pareti di fondo, e dal poco che ne resta nella parete dell'entrata, imita I'interno della copertura di una capanna Ora questa maniera di copertura 6 in armonia con un altro dato di fatto, da noi studiato nelle tracce che si son conservate delle nuove abitazioni suUe pendici prossime. Abbiamo quivi riconosciuti segni dai quali si deduce che se la nuova casa fu fatta ad imitazione del costume greco [i.e., rectangular], con moltissima proba- bility air impianto non corrispose lo stilo della copertura. I canaletti incassati nella roccia, per deviare le acque piovane dall' interno, provano che il tetto fu compiuto col sistema primitivo, oiofe nel modo con cui formavasi 11 tetto delle oapanne. Che la nostra tomba appartenga al periodo delle nuove case a pianta quadrata, e ne riproduca la forma, si revela chiaramente confrontando la pianta che ne abbiamo riprodotto con la pianta delle case " 2 Cf . Pfuhl, "Zur GesoHchte des Kurvenbaus," Athen. Mitteilungen, XXX (1905), 341 f. • I use the term airium testudinatum to include both this type and the kind with a pyramidal closed roof to which the title is more usually applied. Thk Eablt Etruscan and Roman House 117 roof like that in Montelius, Italie Centrale, PI. 175, No. 14. The two cuts in Martha, L'Art Etrusque, Figs. 192' and 193, illustrate the gradual change of this sort of roof as it adapted itself to an ovoid building. The further change, when rectangular buildings became common, is shown by the urn in Martha, Fig. 196, and a still more developed type when the house begins to evolve a ridge-pole, in an urn at Florence.* The roof now slants in two directions; that is, it is not strictly any longer a tectum testudinatum, but rather what Festus' calls a tectum pedinatum. Often the roof was elaborately decorated with antefixes, sometimes in the shape of human or animal heads.* Terraces and loggias' could further complicate this originally simple type of house. Of the interior of this developed form, we may gain an idea from the main rooms of most Etruscan chamber-tombs. The ceiling of the ordinary type is composed of two inclined planes, rising slightly toward the center where they are joined by a horizontal slab carved to imitate the king-beam. Sometimes an even closer representation in stone of a timber-roof is obtained by imitating the side-rafters also. Almost every Etruscan necropoUs shows instances of this style of roofing. I have noted examples from Bieda,' Bolsena,^ Cervetri,* Capua,' Corneto,^" Musarna," Orvieto,^* Otricoli," Perugia," Poggio Sommavilla" (in the Sabine country, not far from Civit^ Castellana), Veil," and Vulci." Many more might be cited. Another common form of roofing for the central chamber is an imitation in stone of wood-coffering. A coffer or caisson consists of a series of rectangles, each within the last, so that finally only a 1 Fig. 192= Montelius, PI. 175, 15. ' Milani, Museo Arch, di Fireme, II, PI. 49, 1 = Martha, Fig. 197. » Festus 260-61, ed. Thewrewk de Ponor. • Cf. Montelius, Italie Centrale, PI. 238, 8, and Milani, II, PI. 48, 1. ' Milani, Mxtseo Arch, di Firenze, II, PI. 49, 1 = Martha, Fig. 197. • N.S. (1877), 152. " lUd. (1863), 43. TBullelino dell' Istituto (1857), 138. " N.S. (1909), 282. 'Ibid. (1834), 98; (1839), 18. "Korte, Volumnier-Grab, 7. • Ibid. (1872), 44. is Bull. d. 1st. (1837), 211. «» Bomische Mitteilungen (1886), 86 S. " N.S. (1882), 292. ^iBull. d. Isi. (1850), 38. " Gsell, Fouilles de Vulci, 533. For a good description of this sort of roofing see N.S. (1877), 152. 118 Margaret C. Waites square opening is left in the center, small enough to be closed by a board or a tile. I shall discuss this kind of ceiling later. Here, it is sufficient to say that I am aware of no instance where a single caisson is employed by itself for the center of the ceiling in a main tomb-chamber. Not uncommon, on the other hand, are instances where the ceiling of the main room or atrium is covered with repeti- tions of the caisson. These coffers are found in tombs at Chiusi and Corneto,^ Perugia,^ Poggio Prisca' (near Sovana), and Vulci.* In all the tombs so far discussed, there are no traces of a roof- aperture in the central chamber. All in this respect resemble the primitive gathering-room of the Italic hut, and this despite two circumstances which might seem likely to suggest another style. The first of these is the extreme elaborateness of some of the tombs. Obviously they are meant to imitate houses of no mean extent, where, one might think, there were so many side-chambers for the life of the family in inclement weather that the central hall might conveniently have been lighted by a large aperture in the roof. For example, the sepulcher of the Volumnii at Perugia contains nine rooms besides the central hall, and some of the tombs at Vulci (see for instance Gsell, Fig. 82) are similarly complicated and display side-chambers larger in proportion to the "atrium." Then too it is often asserted that the Tuscan atrium originated under crowded conditions where houses stood so close to each other that it was not convenient to conduct water by outside gutters from the roof. When, therefore, as at Orvieto, we find adjoining sepul- chers forining parallel and cross-streets and united into insulae, we expect imitations of the Tuscan atrium. Instead, " the slabs which compose the inner walls, from the fourth row upward, constantly converge until they are united by another row of slabs which, fitted like a hinge into the tufa blocks of the two opposite sides, forms, as it were, the key of this false vault."* There are, however, a few apparent instances of a main hall with an aperture in the middle of the roof, and these I shall discuss with care. ' So, for example, at Chiusi (see Annali delV Istituto [1861], 256) and at Corneto (Martlia, L'Art Etrusgue, Fig. 146). See also Martha, Fig. 123. This sort of ceiling is of course extremely common in Greece; cf. Durm, Baukunst der Griechen', 179 ff. » K6rte, op. ei«. « AnnaZi (1843), 225, * Gsell, op. ei«. 'AnnaZi (1877), 101. The Early Etbuscan and Roman House 119 III The first is mentioned by Dennis, Cities and Cemeteries, I, 392. Of the tomb near Corneto called La Mercareccia, Dennis says : The ceiling of this tomb is hewn into the form of a trapezium, with beams on each of its sides, sloping off from the center, which is occupied by a square aperture, tapering up like a funnel through the rock for twenty feet, till it opens in a round hole in the siuf ace of the plain above. In the sides of this chimney or shaft are the usual' niches for the feet and hands It is worthy of remark that in its roof this tomb, which is imique in this particular, represents that sort of cavaedium which Vitruvius terms displuviatum. .... It may be, however, that this opening represents — ^what it more strictly resembles — ^a chimney A steep passage cut in the floor of the tomb leads down to an inner chamber " This inner chamber is represented on Gori's plan^ as almost the size of the outer one. He speaks of the outer cella as a vestibule,' evidently regarding the inner apartment as the main room, a view in which he is followed by Urlichs, writing in Bulletino dell' Istituio (1839), 67, who says of the sepulcher: "It is composed of a corridor, having a hole in the clifl above it, from which by means of a round entrance one reaches the principal grotto." The tomb of the Tarquins at Cervetri is constructed in the same manner, i.e., the inner chamber is on a lower level than the outer; and, in that case, the inner room is plainly the principal one. On the other hand, in a similar tomb depicted by Byres,* the inner section is much smaller. Even, however, if the chamber with the aperture represents the atrium, one hesitates to regard the opening as intended to repre- sent the compluvium, for the simple reason that it is so evidently an entrance. I have said (p. 114) that at times the roof of the atrium testudinatum assumed a pyramidal shape. Now if an opening for descent were cut in the roof of such an atrium, we should have on the inner side an exact reproduction of what we find in the ceiling of the Cometo tomb. On the other hand, we should also have an atrium 1 "Usual" because found in the Civiti Castellana neighborhood with some frequency and occasionally elsewhere; see below, p. 126. 2 Museo Btruaco (1743), III, 90, Class II, PI. 7. *0p. cit., 90: "Exterior haeo Cella, quae vestibuli locum tenere videtur." At times, however, Italian writers apeak of the atrium as a veatibolo; see p. 132. « Hypogaei, Pt. IV, Pis. 1-3. 120 Margahet C. Waites displuviatum, because an atrium displuviatum is nothing more or less than a pyramidal atrium testudinatum with the pyramid trun- kated and pierced by a hole.' The fact that the tomb at Cometo has apparently no other entrance would also argue against the theory that the opening imitates the compluvium.^ Still, if such openings occurred frequently and always in the main room, one would be inclined to believe that besides their actual use as a means of descent peculiar to the tomb, they also imitated an integral part of the dwelling-house. So the wall-paintings in the tombs imitate the decoration of the house, though in subject they are connected with the sepulcher, being probably intended to depict and to assist the happiness of the dead. If my theory that the aperture at Corneto is merely an entrance, not an imitation of the displuviate atrium, be correct, it will gain support from an instance where the non-pyramidal roof, the kind found in the atrium pectinatum, is also provided with an entrance- shaft. For this I turn to the tomb of the Tarquins at Cervetri.' Here the roof of the main chamber, thirty-five feet square, shows two king-beams and rafters connecting them with the sides of the room. Two square pillars support the roof and between them is a shaft cut through to the plain above where it was covered by slabs. Dennis remarks:^ "The shaft was either used as an entrance after the doorway had been closed, by means of niches cut for the feet and hands; or may have served, by the removal of the covering slabs, to ventilate the sepulcher, in preparation for the annual parentalia." In almost all handbooks on Roman private life appears as an example of an atrium displuviatum an urn from Chiusi* cited by ' Daremberg and Saglio, s.v. "tectum," 63. ' "In the Etruscan cemetery of Certosa near Bologna, the graves have no side- entrances, but are large pits, into which one has to descend from above." — Sandys, Companion to Latin Studies, 24. Bartoli (Gli Antichi Sepolcri, PI. 50) gives an illustra- tion of a large round subterranean chamber discovered on the Aventine in 1692, into which one descended by a perpendicular shaft. Bjnres mentions another tomb at Corneto, the Grotta Intagliata, which was entered only from the roof. Of this roof he gives no illustration but as Dennis calls La Mercareccia unique, I judge that it was not of the displuviate type. Dennis mentions sim^ilar tom^bs at Ferento (Cities and Cemeteries, I, 392). » Dennis, I, 242; Martha, Fig. 149. « I, 245». ' Fig. 1, below. The Early Ethuscan and Roman House 121 Martha' as at present in the museum at Florence. As represented in Martha's illustration, the roof has the form of a trunkated pyramid and there is a rectangular opening at the top. It is hard to see how the safety of the contents could have been secured in such an urn, and therefore, before visiting the Etruscan Museum at Florence in search of the original, I con- jectured that the shape of a lost cover might be estab- lished by comparison with another urn in the same mu- fig. i.— Hut-um from cuusi (Martha, l'ah seum.= The top-piece here is -«"•-«-• ^^s- 198). a little curving roof such as would fit exactly on the aperture of the Chiusi urn and might easily, if movable, have been lost. This conjecture received interesting confirmation when I later visited Florence, and a lengthy search proved that the original of Martha's Fig. 198 is not in the Etruscan Museum. According to Milani, the director, who with other oflEicials was most courteous in rendering me assistance, it has never been there. Montelius (Italie Centrale, PI. 226, 5) gives an illustration of it and in the provisional explanation of these plates says: "Pis. 226-231. D'apr^s les originaux dans les Musses de Chiusi, de Firenze et de Londres (PI. 227, Fig. 8)." The urn in question is not in Florence; neither, according to Milani and my own observation, is it in the museum at Chiusi; and apparently only PI. 227, Fig. 8, is referred to by Montelius as in London. He was probably misled by » Martha, Fig. 198= Montelius, Italie Centrale, PI. 226, 5. ' See Martha, Fig. 129, and Milani, Museo Archeologico di Firenze, II, PI. 49, 2. • Fig. 2, above. Pig. 2. — Hut-iim in the Museum at Flor- ence (Martha, L'Art Strusque, Fig. 129). 122 Margaret C. Waites Martha's plate and erroneous reference. Dennis reproduces the urn in his description of Chiusi, but does not say he saw it. As far as I have been able to discover, the only authority who actually saw the urn is Braun who, writing in Bull, dell' 1st. (1840), 150, says it was then at Chiusi and adds: "It forms a house of Etruscan architec- ture, rising on a high quadrilateral foundation and covered by a roof with an enormous projection. On the top were originally placed other architectural features which have now entirely disappeared."^ It should of course be now impossible to use this example as an argu- ment for the antiquity of the Tuscan atrium. These three are the only certain instances I have found of aper- tures or apparent apertures, such as might suggest compluviums, in the roofs of Etruscan atriums.^ The last case may be discounted, and in treating the other two I have tried to show that the opening was in all probability not intended to represent the compluvium. Such apertures may have been used for airing the tomb, for clearing out debris, or for entrance. In all cases of the kind, one must remem- ber that, if the hole imitated the compluvium of an atrium, the impluvium cavity would probably be represented below. In a main room, surely some permanent basin must always have caught the rainwater when an opening existed in the roof. In the roofs of all these tombs, then, one sees prevailing the type of the atrium testudinatum which in a very few examples seems on its way to develop into the atrium displvmatum. Even in these cases such apertures as exist were always intended for some practical 1 "Forma una casa d'architettura tosca, ergendosi sopra alto quadrilatero fonda- mento coperta da un tetto che smisuratamente sporge in f uori. In cima trovaronsi origi- nariamente collocate alire parti architettoniche che oggi sono perdute senza traccia." ' A few doubtful instances may be mentioned here. In describing the roofs of the tombs of Norchia, Orioli in Annali (1833), 30, speaks of "una spezie di piramide tronca, per rappresentare, io credo, il tetto della casa piovente a quattro acque." Cf. the same authority in Inghirami, Mon. Etr., IV, 199 and PI. 42, 2. Dennis, a careful investigator, could find no traces of this fajade (cf. Cities and Cemeteries, I, 204). N.S. (1904), 388, contains an obscure account by Franci of an Orvietan tomb: " L'interno di m. 2 . 25 di lungh., di m. 1 . 60 di largh. edi m. 1 . 85 di altezza, ha panchine intorno di m. . 60 di largh. em. . 48 di altezza, con un' apertura nella volta leggermente arcuata, di m. 4. 15 X -55, con un battente di m. .20 e chiusa da una lastra di tufo." The battente would seem to preclude the idea of a compluvium in this case. For instances of tombs with non-rectangular apertures which could only have been intended for entering or cleaning the sepulcher, see N.S. (1877), 153, and Bull. deW 1st. (1874), 237. The Early Etruscan and Roman House 123 purpose and probably had no other design. After a careful and extended search, I have found only one recorded trace of a shape suggesting the atrium tuscanicum. Lenoir in Annali (1832), 268, mentions the tombs of Toscanella in the following terms: The upper portion of the grottoes of Toscanella always displays two sloping slabs which are more or less inclined, and the representation of a beam supporting them in the middle Some grottoes, and this occurs only when the plan is a square, show in the carving of the tufa the representation of ike pieces of wood which characterize the construction of the Tuscan Atrium.^ The early date of this reference, its inaccuracy, and the fact that I find it corroborated neither by any later notice in the periodi- cals, nor by writers on the general subject of Etruscan civilization, like Dennis, nor by Campanari in his extensive work on Toscanella,* combine to make me discredit it. IV Pertinent to the discussion are the few sites where actual Etrus- can house-foundations of early date remain. Ancient houses are likely to have a central hearth-hole, and at Falerii traces of this were found.' Such houses could never of course have had a large opening in the roof over the hearth. Sometimes even the draught from the door necessitated the moving of the hearth, for in other houses of the neighborhood* one finds in a back comer a little chimney- hole about 8 cm. in diameter. At Marzabotto^ tiles which seem to have formed a sort of dormer- window have been found. Brizio' regards these as perhaps proving 1 "La partie supferieure des chambres de Toscanella offre toujours deux rampans qui ont plus ou moins de pente, et la r6pr6sentation d'une poutre qui les soutient par le milieu Qudgues caveaux, el cela n'a lieu que lorsgue le plan est carri, offreni par le travail opSrS dans le tuf la riprSsentation des pUces de bois qui compoaent la con- struction de V Atrium toscan." * Tuscania e i suoi monumenti, 2 vols. (1856) . There are hundreds of tombs in the vicinity of Toscanella, but many, even of the most spacious, have been covered up again and I found none which agreed with Lenoir's account. The roofs of some were slightly rounded ; others represented the king-beam by a depression and others showed it projecting from the ceiling. Cerasa Giuseppe, secretary of the sindaco and a man of education, told me that he had discovered many tombs and knew the country well. He drew the shapes which the roofs commonly assumed, but none corresponded to Lenoir's statement. » Monumenti Antichi, IV, Figs. 13 and 14. • "Dintorni di Corchiano"; see ROm. Mitt. (1887), 28. » Mon. Ant., I (1890), 298 ff. ♦ Op. cit., 302. 124 Margaret C. Waites the existence of a second story, but it is equally probable that the main room of these Etruscan houses was lighted, like the hut-urns, by a small window over the door. The houses of this town were grouped in insulae and were of considerable extent.^ In two of these dwellings (Insula Villa, IXa), one entered by a vestibule into a large court which was prob- ably open, as it was paved with flints. In one corner of each house was a well and round the court opened rooms without any regular plan. In front there seem to have been shops, and I would suggest that the open court, which reminds one of the Greek house and also of the Pompeian villa rustica at Bosco Reale, might have been used as a yard where carts and bales were kept. Brizio apparently dates the settlement at Marzabotto from the beginning of the fifth century b.c' If this arrangement was at all a common one, we should expect to find it sometimes imitated in the tombs. Dennis (I, 120) speaks of a spot called Puntone del Ponte about two miles from Corchiano where there is "a singular tomb, with a sort of court in front simk in the rock, and with the remains of a portico." In Vulci and its territory one often discovers tombs of the type called "a cassone" which Gsell' describes as con- taining a "vestibule," sometimes square, but oftener rectangular, and open to the sky. This is termed the cassone.* Into it, generally toward the middle of one of the long sides, opens the corridor by which one enters the tomb, and opposite are grouped one or more rooms. Some- times other rooms are found on the short sides of the cassone; sometimes, but more rarely, they appear on the corridor side. Often there are inner rooms without direct connection with the cassone.^ When one compares the plans of tombs like these and 1 Mon. Ant., loc. cit., Pis. I and V. 2 Op. cit, 326. » Fouilles de Vulci, 431 f. < Fig. 3, above. ' Gsell, op. cit., 433, regards this tjrpe as a little later than the simple corridor tomb without the cassone. The vases found in them would date them, according to him, in the sixth century. They disappear some time in the fifth century and the simple corridor tombs take their place. Fig. 3. — Plan of a tomb ol the cassone-type (Gsell, Fouilles de VuUi, Fig. 45). The Early Etruscan and Roman House 125 some of those at Corneto with the houses of Marzabotto and the Roman houses at Ostia, one is tempted to the very sensible conclusion that the Etruscans and the Romans were bound as little as we to the adoption of a single type of dwelling. In the crowded streets of Vetulonia* only one house, No. 19 on the plan, shows traces of an impluvium, though the city was probably not destroyed till about 74 B.C. V One may picture, then, the main room of an Etruscan habitation as of considerable size (the "atrium" of the tomb of the Tarquins is 35 feet square), with a timbered ceiling through the center of which ran usually the king-beam. Sometimes pillars were used to support the roof. In the center was the hearth and over the door was a low, shuttered window from which the smoke issued. Such a room, if surrounded on three of its sides by inner chambers and forced to derive its light from the main door and the window over it, must have been intolerably dismal, especially if a covered vestibule shut off most of the light from the front door. In the coxmtry and sometimes even in town the problem was solved by building the house round an open court. Most of the tombs, however, imply that the atrimn, not the court, was the center. This was probably a later arrangement, developing under more crowded conditions. In general, there was no large central aperture in the roof of the atrium. Was there no other means of introducing light and no other place for a compluvium ? There must have been, for an important passage in Aulus Gellius (x. 15. 8; cf. Servius on Aen. ii. 57) proves the existence of such an opening in very early times. Gellius says : "Ignem e 'flaminia,' id est flaminis Dialis domo, nisi sacrum efferri ius non est; vinctum, si aedes eius introierit, solvi necessum est et vincula per impluviimi in tegulas subduci atque inde foras in viam dimitti." In the tombs of Falerii, a site where Etruscan and Latin influences met, and so valuable as giving a suggestion of possible Roman usage, Dennis^ found one general idea prevailing, characteristic of the site. Unlike those of Sutri, where the door opens at once into the tomb, it here leads into a » N.S. (1895), 274, and (1898), 82. 1 1, 93. 126 Margaret C. Waites small ante-chamber, seldom as much as five feet square, which has an oblong hole in the ceiling, running up like a chimney to the level of the ground above. .... The chinaney in the ceiling of the antechamber probably served several purposes — ^as a spiramen, or vent-hole, to let off the effluvium of the decaying bodies or burnt ashes — as a means of pouring in libations to the Manes of the dead — and as a mode of entrance on emergency after the doors ±' lU. '±. ■ iiiir;i (JlittlllJJCi. VV1(;11 LUXJl \JX UU.'C KA.iayi.U.V MIX) l/jr^JC V^VJOdH, Fouillea de Vulci, Fig. 85). were closed. That they were used for the latter purpose is evident, for in the sides of these chimneys may be seen small niches, .... manifestly cut for the hands and feet.* On p. 328, Dennis mentions traces of a similar entrance before the Grotta del Tifone at Corneto. For a means of ventilation, for entrance, or for pouring in liba- tions, an opening in the main chamber would have answered better than one in the vestibule. The frequency of the occurrence of such openings would render them useless as secret entrances. Such a fashion seems like an attempt to imitate part of the dwelling-house. 1 Cf. also Mon. Ant., TV, 531 f., 543; and Benedetti, Scavi di Narce, 26 and 42. For a somewhat similar type, evidently influenced, like these, by the well-tombs, cf. N.S. (1900), 563. None of these references is so definite as the one quoted from Dennis. Perhaps some of the remains he saw have since been destroyed. The Early Etruscan and Roman House 127 There are traces, too, of openings in other rooms. N.S. (1884), 219, mentions a tomb near Viterbo on the site of the Etruscan Musama, which contained two spacious chambers. In the center of the nearly level vault of the inner room was an opening 1.80X 0.90 m. which led up to the surface of the ground. Steps for descent were carved in its sides and it may therefore have been intended Via. 5. — ^Plan of tomb (Gsell, Fia. 6. — Plan of tomb (Gsell, FouilUa Fouilles de Vulci, Fig. 84). de Vutci, Fig.79). Rooms I, III, V show In their ceilings Imitations ol the king- beam; IV could not be entered. The ceiling of II contains a single caisson. merely for entrance. Similar doubts must always occur in the case of a real opening. Far better, therefore, than an actual aperture would be a simu- lated one. The original significance of such an imitation would, however, be easily forgotten and it would tend to develop into a mere decoration. Sometimes, as in the tomb at Chiusi pictured in Monumenti delV Istituto, V, PI. XXXII, 4 and 6, one finds in an inner chamber a roof with one plain caisson in the center. Some- times, as at Vulci in Chamber V of Tomb CXXXVI, an oval car- touche appears in the inmost square. Or, as in the Frangois tomb 128 Mabgaeet C. Waites at Chiusi (Mon. dell' 1st., V, PI. XIV, 3), an ornamental design appears in the caisson; or in the cartouche, as in the back chamber of the tomb of the Volumnii at Perugia (see below, p. 130). All these forms of decoration may have originated in an endeavor to imitate a simple roof-opening in a back or side chamber. A single caisson, as I have said,* never occurs in a central chamber. Lastly, in one instance, at Vulci,^ a sort of trunkated pyramid rises from the center of an inner chamber. Here we have an adaptation of the displuviate type, but not in the roof of an atrium.' VI Here again it will be satisfactory if actual houses confirm what we have assumed from the construction of the tombs. Of the houses of the "earliest" period at Pompeii, Fiorelli says: They inclose an area of not less than 87, nor more than 328 square meters, covered by a testudinate roof, under which appears a single atrium .... which sometimes has on one side, more rarely on both sides, two or three smaller rooms. It is always, however, without an impluvium, the roof being entirely closed at the top. And here it is worth while remarking that the impluviums of these atriums are not contemporary with the walls, but con- structed in Nocera stone; from which one may infer that the dark dwelling of the first period was lighted within, at a later epoch, by an opening made in the roof * » See above, p. 118. ' Gsell, Fig. 85=Fig. 4, above. « Frothingham, Roman Cities, PI. XX, gives a cut of this room (II) and of Chamber II of Tomb CXXXIV and calls both " main halls." Why, I fail to see. Cf . the plans here reproduced, Figs. 5 and 6, above. *Gli Scavi di Pompeii dal 1861 al 1872, Introd., p. xii: "Esse circoscrivono un'area, non minore di 87, ah maggiore di 328 metri quadrati coperta da un tetto testudinato, sotto cui trovasi un solo atrio . . . . il quale talvolta ha da una parte due o tre stanze minori, piil raramente da entrambi i lati, ma sempre & privo d'impluvio, essendo il tetto chiuso interamente nel culmine. E qui giovi awertire, ohe gl'impluvii di questi atrii non sono contemporanei delle mura, ma costruiti in pietra di Nocera; da cui s'inferisce che la oasa fuligginosa della prima et^ fu internamente rischiarata da un'apertura praticata nel tetto in epoca piil tarda " Fiorelli's "earliest" period is Mau's second period, usually called the period of the limestone atriums. When the impluviums were added is of course extremely hard to determine. The whole period of the limestone atriums was the time of the Etruscan domination. Architecturally, the Etruscan influence lasted long after Etruscan rule over the city ended toward the close of the fifth century. Mau places the terminus ante quern of the architectural epoch at 200 B.C. For at least half a century before that date, however, Greek fashions must have been growing more and more common and with these Greek fashions, as I hope to show, the introduction of the impluvium was con- nected. The Early Etruscan and Roman House 129 Fiorelli finds seventy of these houses in the city. There are, more- over, a few houses in Pompeii which still represent the earlier conditions. In Mau (ed. 1908, p. 362) mention is made of a house Fig. 7. — ^Plan ol house without compluvlum (Mau, Pompeii, 2d German edition, Fig. 192). without a compluvium, of which the plan is here reproduced.^ e rep- resents the atrium, which had no compluvium and was originally lighted by a window and door at the back. These gave on an open court including all the space now occupied by / and g. Opposite the court was another small division of the house, containing k and i. This reminds one of the arrangements at Marzabotto. The kitchen was at one time in i and the atrium was then probably used as a living-room. Ultimately, the atrium apparently lost its impor- tance as the principal room of the family and the kitchen-hearth was moved into one comer of it. In one half of the open court was built a room (g) which perhaps became the most important apartment of the house. To secure light to the atrium, the family, instead of con- structing a compluvium, allowed the other part of the court (/) to remain open as a light-well. In one corner, a cistern (2) collected the rainwater. Overbeck* mentions two more small houses whose builders at- tacked the problem of lighting in a somewhat similar way. Fig. 8, » See Fig. 7, above. « Pompeii, 270 ff. Pig. 8. — Plan of house with com- pluvium back of the atrium (Overbeck, Pompeii, Fig. 148). 130 Margaret C. Waites a house which is assigned to the earliest period of the Roman colony (80 B.C.), shows a compluvium in the very back of the atrium (2), whose roof has no opening in the center. The house in Fig. 9 did not need a light-shaft be- cause rooms 3, 4, and 6 looked directly into the gardens of Sallust and could be illuminated from there. It sloped toward the back until at 8 we find a space, half garden, half kitchen, with a compluvium (9) in one corner which furnished light to the dark passage (5). FiQ. 9. — Plan of house with compluvium at the rear (Overbeck, Pompeii, Fig. 147). VII Does the history of the light-wells end here ? We have seen that the Etruscan tombs and the existing Pompeian houses imply nothing like the rigidity of plan which handbooks would have us assume as characteristic of the Roman house. Nevertheless, some sepulchers reproduce the main part^ of the conventional Roman dwelling, and of these sepulchers perhaps the most noted is the tomb of the Volumnii at Perugia. Here we have atrium, alae, tablinum, and side-chambers. On the inner side of the entrance-wall, a round shield is represented on the gable and a corresponding one on the back wall of the atrium over the entrance to the tablinum. May not such a decoration have its origin in the disks which close the gable-windows of the hut- urns ? The tablinum, as I have remarked above (p. 128), is decorated with a coffer, in the center of which is a Medusa-head. In each of the alae is a corresponding coffer, each decorated with a female head. The possible fimction of the alae as a means of lighting the atrium has long been recognized,^ though not the probability that » Mau, Pompeii (1908), 264 f. Too much emphasis is generally laid on the sup- posed religious importance of the alae. In the first place, they have no fixed position. Sometimes they are found in the middle of the side-walls of the atrium, sometimes there is only one, sometimes none at all. The name was not necessarily restricted to recesses round the atrium, for in two inscriptions (C.I.L., XIV, 4183, and IX, 3523) The Eably Etruscan and Roman House 131 the light in city-houses came usually through an aperture in the roof such as is suggested by the coffers in the tomb of the Volumnii. I would propose a similar function for the tablinum. My idea would be that the Etruscan house, after the hut-stage, was, like the Greek house, built round a large open court. Such houses had no very definite plan; but there was one main room which contained the hearth and was lighted by a door giving on the court and a small window over it. This door could be kept almost constantly open as a street-door could not. As conditions of living grew more congested and safer, so that the older arrangement was no longer possible or necessary, the court disappeared. The main room, now isolated, developed a door toward the street and rooms on the sides. The old opening in the back now formed the entrance into all that was left of the court, a light-shaft furnishing light to the atrium. In this shaft we may recognize the earhest form of the tablinum. 1 Sometimes, a light-shaft might also be constructed in a vestibule or in an ala. The windows, real or simulated, in the walls of the "atria" of Etruscan tombs do not, therefore, necessarily alae are mentioned in connection with a portico. The imagines, when there were any, were far more often to be found in the heart of the ancient house, the atrium, than in the alae. Vitruvius is the only authority who speaks of the alae in connection with the imagines (vi. 3. 6). Writers who mention the imagines incidentally and who are therefore far more valuable witnesses, invariably speak of them as in atrio. So Juvenal viii. 19; Pliny N.H. xxxv. 6; Ovid Fasti i. 591; Amor. i. 8. 65; Martial ii. 90. 6; v. 20. 5-7; Seneca ad Polyb. xiv. 3. ' In discussing the origin of the tablinum, the famous passage from Varro ap.Non. 83 is generally taken as authoritative. This passage, " ad f ocum hieme ac frigoribus cenitabant; aestivo tempore loco propatulo; rure in chorte; in urbe in tabulino, quod maenianum possumus intellegere tabulis fabricatum," certainly implies that the tablinum in city-houses took the place of the courtyard of the villa Tustica. One would like to be sure, first, whether Varro was certain of his facts or whether the words poaaum^ua intellegere do not imply rather a conjecture; and second, whether he is using maenianum in its strict sense, or whether he uses it loosely because his mind is diverted by the verbal connection between tabula and tablinum,. Maeni- anum is too often translated "summer-house" and we are given to understand that it was a kind of rustic lean-to. Of course it can mean nothing but balcony, and, if it is accurately used, we must conceive the tablinum as originally a balcony placed over the door from the atrium into the court, or a veranda opening from the atrium. When, owing to the congestion of neighboring houses, the court disappeared, the balcony would in some degree compensate for its loss, and, when houses in the city became still more crowded, the loggia would develop into a light-shaft. At times a loggia could be inclosed by shutters to form a second-story room. Recently a "sort of impluvium pensile" has been discovered in a balcony. See Classical Journal, IX, 105. 132 Mabgaret C. Waites imply that the inner rooms are thought of as lighted from the main chamber. Usually the atrium was lighted from the inner rooms. After a time, a change in house-construction was introduced. This dated from the adoption of the Greek peristyle.* Diodorus (v. 40) tells us that the Etruscans realized the advantages of the peristyle, and the fact that their own houses had originally been built round an open court would lead them more readily to adopt it. Then the light-shaft at the back became merely a means of entrance from atrium to peristyle, and the atrium itself, once the main room, became merely a vestibule.'' All that it now needed was light and space. There was no need to avoid draughts, for the hearth had been removed from it. It therefore developed an aperture in the roof of the displuviate type. The change would be easy if people had before been accustomed to an aperture in the roof of an ante-chamber, for this was virtually what the atrium had now become. This fashion probably lasted only a short time and perhaps never penetrated as far south as Campania, for we find no examples of it at Pompeii, though two wall- paintings there are said to represent it.' The discomfort of passers-by would soon make this shape impos- sible and necessitate some less projecting form of roof whence the rainwater was not conducted into the street.^ Sometime between the First and the Second Punic Wars, I should date the invention of the Tuscan atrium, which Vitruvius 1 One may doubt whether ordinary houses in a crowded city like Rome usually possessed a peristyle. It is noticeable that the House of Livia on the Palatine has none. 'Gellius xvi. 5: "Pleraque sunt vocabula, quibus vulgo utimur, neque tamen liquido scimus quid ea proprie atque vere significent .... sicuti est 'vestibulum.' Animadverti enim quosdam hautquaquam indoctos viros opinari 'ves- tibulum' esse partem domus priorem, quam vulgus 'atrium' vocat." Cf. also Paul. Fest. 13. > In the Casa dei Capitelli Dipinti and in the Casa dei Dioscuri. See Dennis, I, 392*. It is hard to identify any such frescoes in the houses themselves, and I have not been able to find copies. * Vitruv. vi. 3. 2 mentions a further disadvantage of this tjrpe. Water was likely to collect and stagnate in the pipes used for conducting it from the roof, a process which rotted the walls. The Eablt Etruscan and Roman House 133 mentions first, not because it was the oldest, but because it, and its developed forms, the tetrastylum and the corinthium, were the types with which he was most familiar.' BocEFOBD College ' For convenience I add here a tentative "genealogy" of the Tuscan atrium: Round hut with alightly vaulted roof (Montelius, Italie Centmle, PI. 175, 17) Ovoid hut whose roof had two slopes Round hut with pyramidal roof (Martha, Fig. 193) (D. and S., «.r. "tectum," fig. 6769) Rectangular hut whose roof had two Rectangular hut with pyramidal roof slopes (atrium pectinatum.) (atrium teatudinatum) Atrium displwoiatum Atrium tuscanicwm