fyxmll Winivmii^ Jitetg 
 
 THE GIFT OF 
 
 ..W.C(./^.^v>»jl_aaAA_ ./■,«vrt_-^^=isi.A— bcan/i- 
 
 ..A.....';^.^..o.2-..r^..4> ^ /x ./..<..o. 
 
Cornell University Library 
 NA 323.V21 
 
 The Atrium Vestae/ 
 
Cornell University 
 Library 
 
 The original of this book is in 
 the Cornell University Library. 
 
 There are no known copyright restrictions in 
 the United States on the use of the text. 
 
 http://www.archive.org/details/cu31924008719399 
 
,7 %^^■\ 
 
 THE ATRIUM VESTAE 
 
 BY 
 
 ESTHER BOISE VAN DEMAN 
 
FRONTISPIECE. 
 
THE ATRIUM VESTAE 
 
 BY 
 
 ESTHER BOISE VAN DEMAN 
 
 WASHINGTON, D. C. 
 Published by the Carnegie Institution of Washington 
 
 1909 ,-' , 
 
/\ ^H'^^']Q 
 
 CARNEGIE INSTITUTION OF WASHINGTON 
 
 Publication No. 108 
 
 PRESS OF 
 
 THE Isaac H. Blanqhard Co. 
 
 NEW YORK 
 
PREFACE 
 
 I take this opportunity to express my gratitude to all those, both 
 in America and in Rome, who have so kindly assisted me in the 
 prosecution of this investigation. 
 
 I desire especially to thank the officers of the American School 
 of Classical Studies in Rome for the many courtesies shown me as 
 a research fellow of the School, and the Carnegie Institution of 
 Washington for the generous aid by which the publication of this 
 work has been made possible. 
 
 Esther Boise Van Deman. 
 Rome, June, 1909. 
 
CONTENTS 
 
 Page 
 
 Introduction vii 
 
 I. Historical Introduction ... i-g 
 
 Excavation of the Atrium ... i 
 
 Previous Plans and Opinions . . .... .... I 
 
 Purpose of the Present Work 2 
 
 PeriodsintheDevelopmentof the Atrium ^ 
 
 The Imperial Atrium of the First Period ... 4 
 
 The Atrium of the Second Period .... 6 
 
 The Atrium of the Third Period 6 
 
 The Atrium of the Fourth Period 7 
 
 The Atrium of the Fifth Period .... . .... g 
 
 II. The Republican Atrium Vestae ... . . 9-14 
 
 The Original Atrium Vestae . . . 9 
 
 The Republican Atrium Vestae . ... 11 
 
 Remains of the Republican Atrium 12 
 
 The Domus vestalium 12 
 
 The Domus puhlica 13 
 
 Construction and Materials ... 14 
 
 m. The Imperial Atrium of the First Period . 15-20 
 
 General Description 15 
 
 Plan and Arrangement of the Interior ... 16 
 
 The Republican Altar ... .... 19 
 
 Architectural Details and Construction 19 
 
 IV. The Imperial Atrium of the Second Period 21-28 
 
 General Description 21 
 
 Plan and Arrangement of the Interior 22 
 
 The Sacellum Larum 25 
 
 The Cult Rooms 26 
 
 Architectural Details and Construction . 27 
 
 V. The Imperial Atrium of the Third Period . . 29-34 
 
 The Rooms on the East . 29 
 
 Arrangement and Description of the Rooms . 30 
 Architectural Details . . .32 
 
 The Rooms on the South ... . 32 
 
 Arrangement and Description of the Rooms . .... 32 
 
 Architectural Details . . 33 
 
 Construction and Materials 33 
 
 VI. The Imperial Atrium of the Fourth Period ... .... 3S"42 
 
 General Description ... . ... 35 
 
 Arrangement and Description of the Rooms 35 
 
 The Group of Rooms on the North . . 36 
 
 The Group of Rooms on the South . . 37 
 
 The Meaanino • • 3* 
 
 Purpose of the Rooms . . 3* 
 
 The Shops 4° 
 
 Architectural Details . 4° 
 
 The Upper Stories ... ... • • • 4' 
 
 Construction and Materials . 4* 
 
 VII. The Imperial Atrium of the Fifth Period . . 43-47 
 
 General Description ... ... ... 43 
 
 Plan of the Interior and Description of Changes 43 
 
 Architectural Details and Construction • ■ ■ 45 
 
 Later History of the Atrium ... 45 
 
 V 
 
ILLUSTRATIONS 
 
 Frontispiece. The Atrium Vestae, seen from the East. Facing Page 
 
 Plate I. The Atrium Vestae in 1903 ^ 
 
 II. Fig. 1, TheLevelsof the Atrium. Fig.2, Brick Facing of the Periods of the Flavians and of 
 
 Septimius Severus 4 
 
 in. Fig. I, Remains of the DomusVestalium. Fig. 2, The Earlier Imperial Atrium . . . 12 
 
 IV. Fig. I, Foundation of the Walls of the Imperial Atrium of the First Period. Fig. 2, The 
 
 Republican Altar 16 
 
 V. Fig. 1, The Rooms for the Use of the Cult. Fig. 2, The North Side of the Atrium ... .24 
 VI. Fig. 1, The Rooms on the South, built by Hadrian. Fig.2, Walls of the Third and Fourth 
 
 Periods . . . . . . . . 29 
 
 VII. Fig. 1, The Penus. Fig. 2, The Shrine of the Penates . . 37 
 
 VIII. Fig. I, The Street on the North of the Atrium. Fig. 2, Remains of the Upper Stories .... 40 
 
 IX. Fig. i, The Street East of the Atrium. Fig. 2, The So-called Ovens 44 
 
 X. Fig. I, Arches on the Nova Via of the Time of Septimius Severus. Fig. 1, Stairs of a Late 
 
 Period 46 
 
 Flans A to F. Showing walls of various periods. 
 
 VII 
 
INTRODUCTION 
 
 The House of the Vestals has presented, since its excavation, many 
 problems of great interest to the student of Roman topography and Roman 
 architecture. It was clearly seen that the magnificent building was not 
 the work of one period alone, but that it had been enlarged and beautified 
 in succeeding epochs. The later builders, however, did their work in such 
 a way as to fit it as far as possible into that of their predecessors, and 
 thus produced the impression of a uniform structure. Consequently it is 
 often very difficult to distinguish the different periods. 
 
 Valuable monographs on the Atrium Vestae have been published by 
 Lanciani (1884), Jordan (1886), and Auer (1888); but since their time 
 new researches in the field of Roman architecture and methods of con- 
 struction have thrown additional light upon several questions, while our 
 material for the study of the building has been essentially increased by the 
 most recent excavations (1900-1901). For these reasons a new investiga- 
 tion of the Atrium Vestae had become a necessity. 
 
 In the following pages Professor Van Deman undertakes this task with 
 great energy and with an accurate knowledge of the situation. The House 
 of the Vestals is taken up in her monograph not as an isolated problem, 
 but in connection with a thorough and extensive study of Roman brick- 
 work. As a result of these studies the author has been able to reconstruct 
 the history of the Atrium Vestae in the first and second centuries A. D. 
 For the history of the republican building also, the remains of which are 
 deeply buried under the imperial Atrium and are sketched for the first time 
 in Professor Van Deman's work, valuable suggestions are given. Although 
 some of the author's statements may be subjected to criticism and even 
 corrected by later researches, her work marks a decided advance in the 
 investigation of one of the most interesting monuments of the Roman 
 Forum, and is besides a valuable contribution to the history of the architec- 
 ture of the Romans and to our knowledge of their methods of construction. 
 
 Rome, July, 1909. CH. HULSEN. 
 
THE ATRIUM VESTAE 
 
 BY 
 
 ESTHER BOISE VAN DEMAN 
 
I. 
 
 HISTORICAL INTRODUCTION. 
 
 Excavation of the Atrium: In October, 1883/ in the course of the exca- 
 vations east of the Forum at the foot of the Palatine, there was discovered 
 near the temple of Vesta a building which, from the inscriptions and statues 
 found in and near it, was recognized at once as the house of the Vestal 
 Virgins, the Atrium Vestae.^ During the next three months the whole of the 
 imperial Atrium, with the exception of the rooms on the southwest, was 
 excavated. In 1899 the work, which had been left unfinished in 1884, was 
 resumed, and in 1901, after the demolition of the church of S. Maria Libera- 
 trice, the remaining rooms on the south and west, which before that time 
 had been inaccessible, were uncovered.' The excavations were carried below 
 the level of the imperial period and the scanty remains of the earlier republican 
 Atrium^ were laid bare, wherever this was not rendered impossible by the 
 presence of later structures. During the following two years the excavations 
 in the Atrium and in the buildings adjacent to it were brought to completion. 
 
 Previous Plans and Opinions: Of the results of the excavations of 1883- 
 84,, the first authoritative accounts published were those of Lanciani^ and 
 Jordan.* Their plans of the Atrium' differed but little and were in their 
 main features correct; they were in agreement also concerning the history 
 of the building in considering it an architectural unit and the work of a single 
 period. But in their opinions regarding the specific period to which its 
 construction was to be assigned, they disagreed widely. Jordan, basing 
 his conclusions on the presumable date of the inscription over the cedicula^ 
 and the dates of a number of brick-stamps,' held that the building was the 
 work of Hadrian*". The cedicula is, however, as Lanciani pointed out," not 
 structurally united to the Atrium and need not therefore be of the same 
 
 ' For discussion of the exact date, see Not. d. Scavi, * See plan A, walls indicated in red. 
 
 1883, 371, 470, n. 1. Jordan, Der Tempel ^ JVo(. <i. 5mW, 1883, 434!!. C f. Ruins and Excava- 
 
 der Vesta, 5, n. 5. "'°«^' 4°. ^2.611. 
 
 2 In the fifteenth century, in the search for marbles » Bull, dell' Inst., 1884, 88ff . C/. Der Tempel der 
 
 to supply the limekilns, a number of statues Vesta, z^B. 
 
 with inscribed bases were brought to light, 'Lanciani, Not. d. Scavi, 1883, plate xxn. Cf. 
 
 leading to the supposition that the site was Ruins and Excavations, figs. 72 and 92. 
 
 that of the burial-place of the Vestals. Jordan, Der Tempel der Vesta, plate i. 
 
 'Not. d. Scavi, 1899, 325-333; 19°°' 159-19' • ' JorAm, Der Tempel der Vesta, 27 et al. 
 
 Bull, comun., 1899, 253-256; 1902, 30; "Jordan, /. c, 2.%S. 
 
 1001, 70-78. Huelsen, Roem. Mitlh., xvii, '"Jordan, /. c, 28. 
 
 •^gf ' "£«//. dell' Inst., 1884, 149. 
 
 1 
 
2 THE ATRIUM VEST^. 
 
 period. Moreover, the larger number of stamped bricks were either not 
 found in structural parts of the building or are for various reasons open to 
 doubt in regard to their dates.* Lanciani,^ agreeing with Jordan as to the 
 architectural unity of the building, assigned the whole to the period of 
 Septimius Severus. His conclusions were based upon the general style of 
 architecture in the parts best preserved, and upon the presence, in certain 
 walls, of brickwork undoubtedly belonging to the time of Severus. In 1888 
 Auer, a practical architect, after an independent study of the building, 
 published a brief but suggestive discussion of its history,' with a new plan 
 of the group of rooms on the east.* His conclusions^ differed radically from 
 those of the earlier writers. The Atrium, or rather the part of it then exca- 
 vated, was not, he held, the work of a single builder or period, as Jordan and 
 Lanciani had maintained, but was composed of three distinct units, which 
 were to be assigned to as many periods.' The group of rooms on the east,' 
 the center of which is the large room called by Jordan' the tablinum, which 
 was rightly regarded by Auer as a single structure, was the earliest part 
 of the building and belonged to the period immediately following the fire of 
 Nero.' The group of rooms on the south along the Nova Via, less symmet- 
 rical in plan than the group on the east, was wholly the work of Hadrian.'" 
 The rooms on the north, which were more difficult to identify, since they 
 were only partly excavated, were of a much later period, possibly later than 
 the time of Diocletian." The rooms on the west were not yet excavated 
 at the time of the publication of Auer's work. 
 
 Purpose of the Present Work: The conclusions of Auer as here given 
 were generally accepted as final, when the results of the excavations which 
 were carried on in 1901-03'^ led me to a careful study of the plans of the 
 Atrium, not only those mentioned above but those which have been pub- 
 lished more recently," and of the views advanced concerning its architectural 
 history. After a futile attempt to adapt the plans in detail to the walls as 
 
 > Jordan in his discussion refers to seventy stamps. * Auer, /. c, 20, Note Richter's error (/. c.) in 
 
 Over forty of these are valueless as direct quoting this date, 
 
 evidence, by reason of uncertainty concern- '"Auer, /. c, 21. 
 
 ing their original place in the walls or their "L. c. Middleton {Rtm. of Anc. Rome, i, 309) 
 
 date. The evidence of none of those which accepted in general the view of Jordan, 
 
 are free from doubt is contradictory to the Richter (/. c.) follows that of Auer. Huelsen 
 
 conclusions reached in this discussion. C/. (Huelsen-Carter, The Roman Forum, 206 
 
 Auer, Der Tempel der Vesta, 20. and fig. 125. See also Huelsen, Roem. 
 
 ' Bull. deW Inst., iSS^iHSS. Mitth., 1889, pp. 245-247), adopting the 
 
 ^ Auer, Der Tempel der Vesta, i-io, 20-22. views of Auer in the main, adds to the south 
 
 * Auer, /. c, plate 11. Plate i is a reproduction of rooms those more recently excavated on the 
 
 that published by Lanciani in the Notizie. west, and assigns the rooms on the north to 
 
 ° Auer, /. t., 3, 6-10, 20-21. the period of Septimius Severus. The 
 
 ® A fourth period may be, perhaps, represented by official report of the later excavations is not 
 
 the upper story (Auer, /. c, 8). Richter yet published. 
 
 {Topographie der Stadt Rom, 90) seems so ^Stt plate i. 
 
 to interpret the divisions made by Auer. "Huelsen, Roem. Mitth., xvii, plate i. Vaglieri, 
 
 'Auer, /. c, plate 11. C/. plan C of the present Gli Scavi Recent! nel Foro Romano, 1903, 
 
 work. 15 and 71. Thedenat, Le Forum Romain, 
 
 * Jordan, Der Tempel der Vesta, 36 et at. 1904, 317. 
 
HISTORICAL INTRODUCTION. 3 
 
 I saw and measured them, or to reconcile the facts observed with any of the 
 theories suggested, I decided to make an independent examination of the 
 entire building and of the evidences bearing upon its history. In this exam- 
 ination my purpose was twofold: first, to prepare a more exact plan of the 
 imperial Atrium, so far as it was then feasible, into which the walls discovered 
 since 1889 should be incorporated as soon as they should be made accessible 
 through their official publication;' and, second, to reconstruct the archi- 
 tectural history of the building, including, so far as possible, that of the 
 republican structure beneath it. 
 
 The carrying out of the former purpose was a simple matter, though 
 tedious, consisting merely in the careful measurement of such of the walls 
 now standing as have been published. In the plans based upon these meas- 
 urements, which are here presented,^ a number of errors in the earlier plans 
 have been corrected and some details of importance, as I hope, added. 
 
 The reconstruction of the architectural history of the Atrium was, how- 
 ever, less simple, involving, as it did, the determination of the structural 
 units composing the building and their chronological relation to one another, 
 as well as of the periods to which they are to be assigned. For the determi- 
 nation of the various units and their relation to one another, a careful exam- 
 ination was made of the building as a whole, as well as of the individual walls 
 of the various parts; in this examination special consideration was given to 
 the following points: 
 
 1. The comparative level of the individual walls in each part and the 
 
 relation, with respect to level, between the various parts.' 
 
 2. The unity, in the several parts, in architectural plan and in structure, 
 
 the latter as shown especially by continuity in brickwork and 
 concrete. 
 
 3. The superimposition of walls of one type upon those of another. 
 
 4. The methods of construction, that is, the thickness of the individual 
 
 walls and the occurrence and frequency, in them, of bonding-courses 
 composed of large square bricks, the tegulce bipedales of Vitruvius;* 
 and where it could be ascertained the depth of the concrete founda- 
 tions and the width of the courses of brick and the layers of mortar.' 
 
 ' By the courtesy of Commendatore Boni, Director ^ In earlier investigations concerning brickwork, 
 
 of the Excavations, I continued my work in the width of a course of bricks and a layer 
 
 the Atrium during the progress of the exca- of mortar together has been regarded as a 
 
 vatioDS. I was not, however, allowed to unit of measurement, or the number of 
 
 measure the new walls. courses of brick to the meter has been 
 
 2 Plans A-F. reckoned. Both of these methods are un- 
 
 ' See plate 11, fig. i . reliable, since with the decrease in the width 
 
 * De Arch., vii, i, 7; 4, 2, The teguJee bipedales of the bricks in the later periods there is a 
 
 Lanciani (R. and E., 47) holds appear first corresponding increase in the thickness of 
 
 in the Pantheon and Mausoleum of Hadrian. the mortar. The width of the two together, 
 
 They are used, however, much earlier, therefore, and the number of courses of 
 
 being found occasionally in the walls of bricks in a meter remain in general the same 
 
 Nero, and regularly in those of Domitian. (see plate it, fig. 2). 
 
4 THE ATRIUM VEST^. 
 
 5. The character of the materials employed, as shown by the size, color, 
 and composition of the bricks and by the color and composition of 
 the mortar. 
 
 For the determination of the specific periods to which the various parts 
 of the Atrium belong, a comparative study was made, especially with regard 
 to the methods of construction and the materials employed, of all the build- 
 ings in and near Rome to which a certain date can be assigned. To this 
 evidence was added that afforded by the literature and coins. The number 
 of brick-stamps accessible to me was not sufficient to warrant their use as 
 evidence,* except in a very limited sense. 
 
 Periods in the Development of the Atrium: As a result of my investigation 
 along the lines just indicated, certain important facts have been established 
 and new conclusions reached concerning the architectural development of 
 the Atrium. The structural units^ of which it is composed, apart from those 
 of the early republican building, the number of which it is impossible to 
 determine accurately, are seven or, possibly, eight. These will be described 
 more fully in connection with the discussion of the architectural details of 
 the several Atria. The stages in the history of the building represented 
 by these seven units are, however, but five in number. While no conclusive 
 evidence remains concerning the exact dates of these various stages, the 
 periods to which they are to be assigned can, as will be seen, be definitely 
 determined in all cases. 
 
 The Imperial Atrium of the First Period: The building whose scanty 
 remains lie a meter below the present level of the Atrium' may be accepted, 
 from its orientation and style of architecture, as the republican Atrium 
 Vestae. More than half a meter above this was erected another building, 
 consisting of a court 45 meters long, which was surrounded on two or, 
 possibly, three sides by a series of lofty rooms.* This later building, which 
 differed in orientation and architecture not only from the earlier Atrium, 
 upon the remains of which it was erected, but also from the other buildings 
 of the precinct contemporaneous with it, belongs, as is evident from its 
 construction, wholly to the imperial period. In the earlier of the two build- 
 ings, the republican Atrium, whose architectural history extends from the 
 early republican or even the regal period to that of the early Empire, many 
 of the walls were restored more than once before their final destruction, and 
 new walls were added, especially in the rooms which belonged at an earlier 
 
 1 The value of brick-stamps in determining the ^ By a structural unit is here meant a building or a 
 date of the structures in which they are found part of a building in which the construction 
 
 has been greatly overestimated, owing to the is identical in type and the walls are con- 
 
 failure to take into account the frequent use tinuous throughout, 
 
 of new material in repairing old walls and ' Plan A, ii-xii. 
 
 of old material in constructing new on?s, * Plan A and pp. i6(f , 
 
E. B V. D. 
 
 Fig, 1. The Levels of the Atrium. 
 
 
 Fig. 2. Brick Facing of the Periods of the Flavians and of Septimius Severus. 
 
HISTORICAL INTRODUCTION. 5 
 
 period to the Domus publico. All of these restorations and additions are, 
 so far as can now be determined,^ antecedent to the fire of Nero. It is 
 certain that at that time the whole precinct was swept by the flames.^ Had 
 the Atrium survived the calamity, there would be traces remaining of the 
 restorations necessitated by it. Since, as we have said, no such traces are 
 to be found, the final destruction of the earlier building can not have been 
 subsequent to that event. The later building also, which was erected in its 
 place — ^which we have called the first imperial Atrium — though it suffered 
 at least twice from fire, shows no evidence of any such complete restoration 
 as would have been necessary had it been built before and passed through 
 the great fire. We may conclude, therefore, that the earlier building, the 
 republican Atrium, was destroyed in the fire of Nero and that the first 
 imperial Atrium, by which it was replaced, was erected at some time sub- 
 sequent to it.' We know that at the time of the death of Galba, in 69 A. D.,^ 
 the precinct of Vesta was recognized as a place of refuge and that the temple 
 and certain of the less important rooms connected with it were already built. 
 At that time, therefore, the Atrium must have been in large part completed, 
 since the continual attendance of the Vestals upon the temple-fire made it 
 necessary that, with the restoration of the regular temple cult, they resume 
 their residence inside the precinct. The half-year between the death of Nero 
 and that of Galba would have been insufficient for the carrying out of the 
 whole work;' the reconstruction of the Atrium as well as of the temple" 
 must, then, have been begun, if not finished, by the former.' The arguments 
 presented are, as it seems to me, sufficient in themselves to warrant the 
 acceptance of the first imperial Atrium as the work of Nero. This conclusion 
 is, however, made more certain by the evidence of the walls themselves, 
 which are identical in material and in methods of construction with those 
 universally recognized as belonging to his time.* 
 
 •The restorations and additions belong in large <Plut., Galba, 27. Tac, Hist., 1, 43; Piio in 
 
 part to tlie Domus puhlka. Of these new adem Vesta pervasit, exceptusque mistri' 
 
 walls but one type is uncertain. Although cordia puhlici servi et contuhernio eius ab- 
 
 no decision can now be reached concerning ditus prolractus Piso in joribus 
 
 its date, in no case need it be held to be templi trucidatus est. 
 
 later than the early part of the reign of Nero. ' The coins of Galba (Cohen, Galba, 309-314, 364, 
 
 For further discussion, see pp. 12-14. 3^7> 4°4) 43^)> ^^ ^'1' ^^ those of Vitellius 
 
 ^ Tac, Ann., xv, 41. (Cohen, Vitellius, 89-91), afford evidence of 
 
 ° As filling for the concrete foundations of this their continuance of the work. 
 
 period, rare marbles have been used in ap- ^ Owing to the importance of the cult, neglect 
 
 preciable quantities. Before their abandon- to rebuild the temple would hardly have 
 
 ment to such a use, these marbles must have escaped the notice of the historians, 
 
 been rendered worthless for other purposes ' H. Dressel {Zeitschr. fiir Numismatik, xxii, 23, 
 
 by the destruction of the earlier buildings n. 3) holds that the temple was only planned 
 
 in which they had been used. Before the by Nero, but built at a later time. The evi- 
 
 rise of the new city of Augustus, imported dence of the coins is not in harmony with 
 
 marbles were almost unknown. After that this conclusion; for, though the variants in 
 
 period no destruction befell the Atrium or type are not numerous, the coins represent 
 
 the buildings adjacent to it until the fire of a number of issues. For example, the three 
 
 Nero. The first imperial Atrium must, gold coins in the British Museum, though 
 
 therefore, have been erected after that of one type, are from three different issues, 
 
 calamity. ' See pp. 19-20. 
 
6 THE ATRIUM WESrjE. 
 
 The Atrium of the Second Period: Within a few years after the erection 
 of the first imperial Atrium, it suffered considerable injury from fire.' The 
 reconstruction following this partial destruction constitutes the second stage 
 in the history of the building.^ From literary evidence' we know that in 
 the reign of Vespasian the Templum Augusti fell a prey to the flames. Since 
 the injury to the Atrium of which we have just spoken was, so far as can be 
 ascertained from the walls now standing,* confined largely to the rooms on 
 the west and southwest,' it is probable that it was the result of this same 
 conflagration, which, entering the building from the direction of the temple 
 of Augustus, swept over this part of it only. From the coins of the Flavian 
 emperors' it is clear that the temple of Vesta was at some time either partly 
 or entirely rebuilt by them. Since there is no reference to any injury to the 
 temple during this period, or to any other calamity befalling this part of the 
 city which might have caused its destruction except that just mentioned, 
 we may safely assume that the rebuilding of the temple indicated by the 
 coins, and the restoration of the Atrium which very naturally accompanied 
 it, were necessitated by the injury wrought by this fire. To this presumptive 
 evidence is added the indisputable evidence aflForded by the masonry of the 
 building, which is of the same type as that of the numerous other buildings 
 of the Flavian emperors, especially of Domitian.' To this time, therefore, 
 we may assign the imperial Atrium of the second period.' 
 
 The Atrium of the Third Period: There is no evidence, either direct or 
 indirect, of any further change in the Atrium until the second century, 
 when with the growing luxury of the times there arose the necessity for a 
 more extensive establishment. To satisfy this necessity it is probable that 
 the group of rooms on the east' was added. At the same time on the south 
 there was built, in the tablinum of the earlier Atrium, a smaller group of 
 rooms."* The Atrium so enlarged we have called the Atrium of the third 
 period. Concerning the exact time at which these additions were made, the 
 proofs are simple and decisive. The new rooms on the south are, as will be 
 
 The irregular manner in which certain of the walls * Coins bearing representations of the temple: 
 
 have been repaired indicates destruction by Cohen, f'es/iasien, 577-581; Tifuj, 347-351; 
 
 fire rather than intentional demolition. For Domiiien, 613-616. Coins referring to the 
 
 the extent of the injury, see plan B and pp. goddess or to the cult: /. c, Vespasiertj 572- 
 
 2ifl. 576; Titus, 340-346; Julie, fille de Titus, 
 
 2 Plan B. J5-18; Domitien, 6ii-6i2. The represen- 
 
 ' Plin., N. H., xn, 94. tation of the temple on these coins differs 
 
 * Cf. plans A and B. For further discussion, see from that on the coins of Nero, suggesting 
 
 p. 21. an entire rebuilding of the temple and not 
 
 ^ The outer walls on the north, which are standing the completion of a structure already begun. 
 
 to a considerable height, show no sign of ' See pp. 27-28. 
 
 restoration at this time. Those on the south- ' Huelsen holds correctly that the Templum Augusti 
 
 east are in part restored. The exact extent as it now stands is the work of Domitian. 
 
 of this restoration can not at present be The restoration of the Atrium by him is 
 
 determined. It is, however, clear that it therefore more probable. 
 
 was by no means as complete as that of the ° Plan C, 29-39. 
 
 western part of the building. '"Plan C, 13, a, b, c, d. 
 
HISTORICAL INTRODUCTION. 7 
 
 seen,* later in construction than the rooms adjacent to them. They must, 
 therefore, belong to a period subsequent to that of Nero and of the Flavians, 
 to which these rooms have been assigned. The rooms on the east are 
 identical in type of construction with those on the south.^ The number of 
 coins of the time of Hadrian^ which refer to the cult, as well as the well- 
 known activity of that ruler in building, lead to the assumption that the 
 enlargement of the Atrium is to be ascribed to him rather than to any 
 other of the post-Flavian emperors. With this conclusion the walls are in 
 harmony, agreeing even in many unessential details with those of the more 
 important monuments of Hadrian, such as the Mausoleum and the Pan- 
 theon.^ The new groups of rooms on the east and on the south are, there- 
 fore, of his time. 
 
 The Atrium of the Fourth Period: In the previous period there had been 
 left, on either side of the garden^ on the east, spaces not unlike the al<s of the 
 private house. By filling in these spaces the number of rooms was much 
 increased and the newer part of the Atrium on the east was more closely 
 connected with the earlier building," although the court on the west and the 
 garden were not united to form the present large court until a later period.' 
 Over the eastern half, at least, of the Atrium, when so enlarged, upper stories' 
 were built, or possibly, in the case of the rooms already existing, rebuilt. 
 These alterations and additions mark the fourth stage in the growth of the 
 building." That we are here dealing with a distinct structure and one later 
 in period than those already discussed is certain. It is evident, in the first 
 place, from the lack of agreement in architectural plan and in type of con- 
 struction, as well as from the loose mode of attachment between their walls, 
 that the rooms which are here under discussion are structurally distinct from 
 those both of the first and second periods on the west and of the third period 
 on the east."* That they are not only distinct from, but later than the rooms 
 of the other periods is even more evident. The concrete foundation of the 
 front wall of the east rooms projects more than half a meter beyond the wall 
 itself. On the north side of the court the later wall, the foundation of which 
 is a meter lower than that adjoining it, has been built, in a bungling fashion, 
 against and on top of this projecting foundation." In the same manner, 
 on the south a shelf which extended along the front wall of the earlier rooms 
 has been utilized in the building of the new walls.'^ The walls of the upper 
 stories also,'' which are united structurally with the new rooms of the lower 
 story,are built on top of those of the east rooms,"or are, as on the side towards 
 
 ' See p. 32. 'Seep. 43. 
 
 2 See pp. 32f . ' Plan Fa. 
 
 ' Cohen, Adrien, 1450; Sabine, 64-68, 78-87. The "Plans D and Fa. 
 
 coins of the time of Trajan are few (Cohen, '"Auer (/. c, p. 8) noted certain of these facts and 
 Trajan, 644-645; Ploune, lo-n). The called attention to their significance, 
 
 brickwork also does not belong to his period. "See p. 36 and plate vi, fig. i. 
 
 •Seep. 33. "Seep. 36. 
 
 ' See plan C, t. "Plan F. 
 
 " Plan D. "See p. 42. 
 
8 THE ATRIUM VESTwffi. 
 
 the Palatine, pushed out above and beyond them. The structural unit, 
 therefore, of which the walls of the upper stories form a part, must be of a 
 period later than that of these rooms, which are of the time of Hadrian. The 
 immediate successors of Hadrian, in fact, the Antonines, especially Lucilla, 
 were liberal benefactors of the Vestals.^ With the architecture of their time, 
 moreover, that of the new building is in agreement.^ We may, therefore, 
 conclude that the additions to the Atrium by which the eastern half of the 
 building was completed are to be assigned to the period of the Antonines 
 and are probably in a large part the work of Lucilla. 
 
 The Atrium of the Fifth Period: During the reign of Commodus the 
 precinct of Vesta suffered again from fire. In this case the fire entered the 
 Atrium from the side towards the temple, which was, as we know, almost 
 totally destroyed by it. The injury afi^ected most seriously, therefore, the 
 western and northwestern rooms, the inner walls of which were burned 
 almost to the ground.' The rooms on the east and south, however, escaped 
 injury. The restoration following this partial destruction marks the last 
 fixed stage in the development of the Atrium.^ In connection with this 
 restoration the Atrium-court on the west was extended to the east by the 
 destruction of the walls separating it from the earlier garden' and by the 
 lowering of the level of the garden to that of the court. Discussion of the 
 exact period at which this restoration took place is rendered needless by the 
 peculiar nature of the brickwork, which is that accepted without question 
 as of the time of Septimius Severus. Julia Domna, the wife of Septimius 
 and a well-known patroness of the Vestals, rebuilt the temple." It is likely 
 therefore, that the restoration of the Atrium was also her work. 
 
 After the time of Julia Domna, though the Atrium shared in the vicissitudes 
 which befell the buildings in its vicinity, the restorations and additions were 
 of comparatively small importance and can not be assigned to any certain 
 period. 
 
 ^ The coins of the period are numerous- Cohen * Plan E. 
 
 Faustine Mere, 285-293, 318; Marc Aurele ^ See p. 43. The extension of the court may be of 
 
 et Lucius Verus, 2; Faustine Jeune, 284- the preceding period, but the weight of 
 
 286; Lucille, 92; Crispine, 45. It is as- evidence is rather in favor of the view here 
 
 sumed that the famous medallion of Lucilla given. 
 
 refers to a restoration of the temple. It is ^ This is clear from the coins of the period, as well 
 
 unhkely that the temple was restored by as from other testimony. Coins represent- 
 
 her. It is possible that the medallion was ing the temple: Cohen, Julie Domne, 232- 
 
 struck instead, in token of assistance given 244. Coins representing Vesta: I.e., 220- 
 
 the Vestals in the completion of the Atrium. 231, 245-248. It is likely that the temple at 
 
 ^ Unfortunately very little masonry of the time of least was not completed, since there are 
 
 the Antonines remains. coins of Caracalla (/. c, Caracalla, 249-251) 
 
 ^ See plan £. on which it appears. 
 
II. 
 
 THE REPUBLICAN ATRIUM VEST^. 
 
 Although not connected directly with the main subject under discussion, 
 the imperial Atrium Vestae, the lack of definiteness in the use, among ancient 
 and modern writers alike, of the term Atrium Vestae has led me to discuss 
 briefly the earlier use of the name and its gradual restriction to that build- 
 ing to which it is commonly applied. 
 
 The Original Atrium Vesta: During the earlier period of its existence 
 there were no distinct buildings inside the precinct of Vesta,^ but the various 
 parts, united more or less closely among themselves, formed a single complex 
 structure.^ In this group of buildings were included the king's house proper 
 on the north and the Domus Vestalium, or private rooms of the Vestals, on 
 the south, the small temple, or more properly shrine, of Vesta in the center, 
 and the group of rooms on the east, known later as the Domus puhlica; along 
 the Nova Via at the foot of the Palatine lay the lucus Festce? The center of 
 this group of buildings was the open court in which stood, as in the private 
 house, the shrine of Vesta. This court was in form and orientation a templum, 
 which at an early period was probably inaugurated, though later this was not 
 feasible, owing to the nature of the cult of which the temple was the center. 
 The walls on the north and on the south are in part preserved. On the west 
 it was probably bounded by the outer wall of the precinct.* Towards the 
 east, that is, along the front of the templum,^ the line bounding it was retained, 
 even after the change in orientation of the surrounding buildings, in a raised 
 edge, or curbing,' in the pavement, which crossed the area diagonally in 
 
 ' This is called by the Italians locus Vesta. Cf. ' Serrius, ad Am., vii, 153: Templum Vesta non 
 
 Maes, Vesta e Vesiali, 157 et al. Marucchi, full augurio consecratum ne illuc conveniret 
 
 Le Forum Romaitif 160 et aJ. As a technical senatus uhi erant virgines. The templum 
 
 designation for the precinct, however, the here is not the ades but the area, or court 
 
 name is not found in ancient writers and just mentioned. The ades, because of its 
 
 possibly had its origin in a mistranslation form, could not be inaugurated, and from 
 
 of Ovid, Trsst., iii, 1, 29: Hie locus est its size and arrangement would have been, 
 
 Vesta, where Vesta is predicative and does in any case, impracticable as a meeting- 
 
 not depend on hcus. place for the Senate. Into this court, in 
 
 ' The road which now divides the temple from the which there was, as is shown on coins, an 
 
 later Regia is not original. altar, Scsevola fled for refuge in 82 B. C. 
 
 ' Cic, de Divin., i, 45. (Cic, de Nat. Deorum, ni, 32, 80. Livy, 
 
 * There are, however, in the rear of the temple Epit., lxxxvi et al.) It is possible that the 
 
 traces of a row of rooms, which may have sacriflcial strata found at the southwest of 
 
 formed the western line of the templum, the temple (JVo/.^.iScaui, 1900, 172) may be 
 
 though the remains which are at present in some way connected with the exauguratio 
 
 visible are too scanty to allow of any certain of this templum. 
 conclusion concerning them. Plan A, Z " Plan A, y. 
 
 9 
 
10 THE ATRIUM VEST^. 
 
 front of the later adicula. The pavement itself, the eastern half of which is 
 raised slightly above the other, like the curbing, still preserves the earlier 
 north and south orientation.* The presence of this square inclosure vpas 
 recognized by Jordan, who, however, considered it a separate temenos, or 
 precinct, and held that its orientation, like that of the temple, was due to 
 Greek influence.^ The rooms on the north side of the court with those on 
 the east' formed the residence of the king, though this was, at least in later 
 times, merely official. The rooms on the south were set aside for the Vestal 
 Virgins, who were simply a part of the king's household, assisting the regina 
 in her duties as mater familias of the state. 
 
 This whole group of buildings, called more commonly the Regia of the 
 kings,^ was the original Atrium Vestae. This conclusion is based upon the 
 following facts. It is evident from the numerous references in literature' 
 that there existed from the earliest times a building or group of buildings 
 inside the precinct which was known as the Atrium Vestae. It is, moreover, 
 implied, or rather assumed as known, by Servius" that this earlier building 
 or group of buildings existed before and was distinct in location from the 
 Atrium of the Empire, with which he was familiar. The original Atrium 
 must have consisted of, or included as its most prominent feature, an open 
 court, which corresponded in general style to the Atrium of the private house, 
 since from this its name is obviously derived.' This court must also have 
 been of a size sufficient for the meetings of the senate,* which numbered 
 during the period with which we are concerned, from one hundred to six 
 hundred members. As shown by recent excavations, the precinct of Vesta 
 contained no court fulfilling these conditions except that inclosing the 
 ades, which has been described above." This court, therefore, or rather 
 the group of buildings of which it was the center and to which the name 
 was easily extended, was, as has been said, the earlier Atrium Vestae. The 
 use of the designation Atrium for such a group of buildings is nor peculiar. 
 The best-known examples are the Atrium Libertatts,^" the Atrium Sutorium,^^ 
 and the Atria Licinia in prima SuhuraP The location of the cedes inside 
 the Atrium finds interesting parallels in the location of the shrine of Minerva 
 
 ' C/. Jordan, /. c, 23, and plan B. ' For examples see Gell., 1, 12; Ovid, /. c, vi, 263; 
 ' Jordan, /.c, 83. Servius, ad Aen., vii, 153; Plin. EpisI, 
 
 'The recognition of these rooms as a separate vii, 19. 
 
 structure took place much later. See p. 11. ^Servius, /. c, Ad Atrium Vesta conveniebat 
 ' The question as to whether the Regia and the {senatus) quod fuerat a temph remotum. 
 
 Atrium Vestae are identical has been much The meaning of this much-disputed passage 
 
 discussed. For principal references, see seems clear from the recent excavations. 
 
 Ambrosch, Studien und Andeutungen, 3zff. ' Jordan, Der Tetnpel der Vesta, ■}&!. Cf. Top., i, 
 
 The identification by Ovid is of especial i, 532, 533, n. 6i. Marquardt, Rom. 
 
 interest. Fasti, vi, 263-264: Hie locus Staatsverw., in, 159. 
 
 qui sustinet Atria Vesta Tunc erat intonsi * See above, u. 6. 
 
 Regia magna Numa. By a confusion of the ' Pp. 9-10. 
 
 two names arose later the designations '"Jordan, Top., 1, 11, 460!. Richter, Top., io8f. 
 
 Atrium Regium (Livy, XXVI, 27, y, xxvii, "Jordan, /. c, i, 11, 452. Richter, /. c, 307. 
 
 II, 16) and Regia Vesta (Orelli, 2353). '^Jordan, /. c, i, 11, 433; in, 331, u. 21. 
 
THE REPUBLICAN ATRIUM VEST^. 11 
 
 inside the Atrium Minervcs, as suggested by Huelsen/ and of the leJes Titi 
 inside the Templum Divorum? The application of the name of Vesta alone 
 to the whole Atrium arose naturally from the prominence of her cult among 
 the sacra of the state which were centered there.' 
 
 The Republican Atrium Vestce: With the gradual breaking up of the 
 simple cult of which the king's house had been the center, and the growing 
 independence of the various priesthoods among which the several religious 
 functions of the king had been divided, the necessity arose for the assignment 
 to them of distinct official residences. At this time it is probable that the 
 parts of the Atrium became independent; for during the later Republic and 
 the early Empire, in place of a single complex structure bearing one name, 
 there were recognized four separate parts with as many distinct names, two 
 of which were, however, those applied earlier to the whole structure. These 
 four parts were: (i) the rooms on the north of the temple area, which at 
 least after their reconstruction by Domitius Calvinus, in 36 B. C, were 
 regarded as a distinct structure, and to which was technically restricted 
 the name of Regia; (2) the cedes (less correctly called the templum) Vestce; 
 (3) the rooms on the south, the Domus Vestalium, to which the name Atrium 
 was probably limited;* and (4) the rooms on the east, the Domus publico, 
 in which the Pontifex Maximus' continued to live until Augustus on his 
 assumption of the priesthood, in 12 B. C.,° transferred the official residence 
 to the Palatine.' The names Atrium Vestae and Regia, however, though 
 technically restricted to the parts named, were used also in the earlier and 
 more general sense.* The lucus Vestce, which covered the space originally 
 between the Atrium and the Palatine, yielded place gradually to the new 
 buildings, though a small part of it remained until a very late period. 
 
 After the destruction of the whole group of buildings in the fire of Nero," 
 the Atrium was rebuilt on a scale commensurate in size with the other 
 buildings of the period and with a different orientation from the earlier 
 buildings.'" In this reolmstruction the Domus publico disappeared wholly. 
 
 • The Roman Forum, 117. ' The rise into prominence of the pontifical power 
 2 Jordan, /. c, ill, 565. was coincident with or followed closely after 
 > A parallel is found in the title of the priestesses. the fall of the kings (Marquardt, Rom. 
 
 Their office was that of the mater familial, Staatsverw., iii, 2355.). The granting of 
 
 but from the prominence of the one cult the Domus puhlica to the Pontifex Maximus 
 
 among the many committed to them arose as his official residence occurred probably 
 
 the title Virgines Vestales. Cf. Pontifices at the same time. 
 
 Vesta. ° See Wissowa, Religion und Kultus der Romer, 69. 
 
 * It is possible that the name Atrium included at ' Dion Cass., liv, 27. 
 
 that time not only the Domus, but also the ' By Cicero (ad Att., x, 3), as well as by other 
 
 area of the temple. The remains (plan A) writers, the name Regia is used for the 
 
 show that a peculiarly close relationship Domus puhlica. Examples occur also of 
 
 existed between the two parts until the the use of Atrium Vestffi for the entire group 
 
 destruction of the whole building in the of buildings, 
 
 first century A. D. After that time the ° See p. 5. 
 
 Atrium was much less closely connected '"See plan A. 
 with the temple. 
 
12 THE ATRIUM VEST^. 
 
 Remains of the Republican Atrium: The remains of the original republican 
 Atrium* as a whole are scanty, though easily distinguished from those of 
 the later buildings on account of their difference in orientation and in con- 
 struction. The remains of that part of the building with which we are more 
 immediately concerned, the early house of the Vestals, to which in 12 B. C. 
 the Domus publico had been added, are even fewer than those of the other 
 parts.^ There are distinguishable, however, even in these scanty remains, 
 three periods of construction.^ The walls belonging to the first of these 
 periods are, like those of the Regia proper, of tufa of two different varieties. 
 Of these the walls of cappellaccio* are somewhat the older, though those of 
 light-yellow tufa may be but a little later. At an early period certain of 
 these walls were restored and new ones added in the harder reddish-brown 
 tufa. In 12 B. C.,^ or a little later, extensive changes were made in the 
 Domus publico both in plan and in type of construction, either to render it 
 more serviceable to the Vestals or in consequence of some partial destruction 
 of the earlier building." At the same time some unimportant changes took 
 place in the Atrium. To this period belong the numerous brick-faced walls' 
 by which the tufa walls of the older building have been replaced and its 
 larger rooms and central court cut into smaller rooms.' 
 
 The Domus Vestalium: The original house of the Vestals occupied the 
 space between the temple area on the north and the lucus Vestce on the south, 
 the precinct of Juturna on the west and the Domus publico on the east. The 
 level is a little more than a meter below that of the later imperial Atrium. 
 The orientation is, like that of the precinct as a whole, north and south. 
 The apportionment of rooms between the Domus publico and the Domus 
 Vestolium is difficult, since "the common wall" of which Dion Cassius' 
 speaks can not be determined definitely. Of that portion of the whole which 
 belonged certainly to the house of the Vestals, the principal parts distin- 
 guishable are a small vestibule-court" and a series of rooms" along two sides 
 of it. The court opened directly from the central area in which the temple*^ 
 stood and was about 20 meters long and 7 or 8 meters wide. It was paved 
 with a mosaic pavement made of a white limestone resembling marmor 
 palombino and silex, into which were set at irregular intervals larger pieces of 
 
 ' See plan A, walls indicated in red. "In 14 B. C. the Basilica Aemilia with the build- 
 2 Plan A, ii-xii. Since the official reports con- ings near it was burned. It is possible that 
 
 cerning the excavations are not yet published, the Atrium may have suffered also. 
 
 no exact plan of the republican walls is ' Plan A, walls outlined in indigo. 
 
 possible. The partial plan given is not ' Middleton recognized three periods of construc- 
 
 intended to be more than a suggestion of tion in the early Atrium. The division 
 
 the main features of the building. made by him (Archaologia, xux, 399ff .) is, 
 
 i The first two periods have not been distinguished however, inaccurate. 
 
 on the plan. * Dion Cass., uv, 27. One of two walls is possible. 
 * An inferior kind of tufa, which is found in the The evidence concerning either of them is 
 
 buildings of the earliest period. not decisive. 
 
 ° The Domus publico was granted to the Vestal '"Plan A, 11. 
 
 Virgins by Augustus in that year. Dion "Plan A, iii-ixand plate in, fig. i. 
 
 Cass., Liv, zy. "Plan A. 
 
PLATE 111. 
 
 E. B. V. D. 
 
 Fig. 1. Remains of the Domus Vestalium. 
 
 The Earlier Imperial Atrium. 
 
THE REPUBLICAN ATRIUM VEST^. 13 
 
 bright-colored marble.' Out of this court, on the south, open five or six 
 small rooms,^ which are not more than 4 meters long and vary from 2.50 to 
 4 meters in width. On the west are other rooms' the dimensions of which 
 can not be ascertained, since they have been almost wholly destroyed by the 
 erection of the later building. In the smaller rooms* are pavements similar to 
 that in the vestibule, but of a finer quality.' Beyond the rooms on the south 
 was an area paved with blocks of cappellaccio, opening probably upon the 
 earlier Nova Via.' In a room, or court, on the southwest was a rectangular 
 basin, ^ resembling the impluvium of the private house. 
 
 The Domus Publico: The extent of the Domus publico is uncertain, but 
 it is probable from the existing remains that it was much larger than the 
 Domus Vestalium. It is likely that it covered almost the entire space between 
 the Sacra Via' and the earlier Nova Via. On the west it was united to the 
 house of the Vestals by "the common wall" of which we have spoken. On 
 the east it may have extended to the precinct which inclosed the ancient altar 
 destroyed later by Nero. It is more probable, however, that the group of 
 rooms about the court on the east" belonged to another house.'" The general 
 plan of the Domus can not be determined. In the center of the earlier 
 building was an open court," of which there remain the bases of two columns 
 and a gutter of travertine, as well as other smaller fragments. Of the rooms 
 opening upon this court but little remains. In the later reconstruction this 
 court was replaced by a number of smaller rooms. Beyond it on the east 
 there were other rooms. The largest of these," in which there is an apse, 
 was possibly a center for some of the various rites over which the Vestals 
 presided. 
 
 In the earlier excavations some fragmentary remains were found of the 
 painted stucco with which the Domus publico was decorated.'' But few 
 traces of this are now visible, though the pieces of marble inserted in the 
 walls to furnish a hold for the stucco are still to be found in many places 
 The style of decoration" is very simple, consisting of panels in plain colors, 
 marked off by narrow lines, in which were circles or simple floral designs. 
 
 ' The pavement in the vestibule of the house of ■ Along the Sacra Via was a row of rooms or shops 
 
 Livia at Prima Porta is of the same type which may have formed a structural part of 
 
 and probably of the same period. Similar the Domus, 
 
 pavements found frequently in Pompeii ° Plan A, xiii. 
 
 belong to the late republican period. '"Behind the room with an apse (Plan A, xi) there is 
 
 2 Plan A, iv-ix, and plate iii, fig. 1 . a heavy wall, which may be the division wall 
 
 ' Plan A, III. between the two houses. If this be so, the 
 
 * Plan A, IV-IX. entire building to the west of this wall must 
 
 ^ The pavement in these rooms resembles that of be included in the house of the Vestals and 
 
 the private house destroyed by Nero in the the Domus puhlica con£ned to the group of 
 
 construction of his Golden House, of which the rooms on the east. 
 
 a small part still remains below the Baths of "Plan A, a. 
 
 Trajan. i^pljn A^ xi. 
 
 ' The course of The Nova Via was originally fur- ''Middleton, Archxologia, xlix, 40of. 
 
 ther towards the north than at present. '*See Middleton, /. c, xlix, plate opposite p. 402. 
 ' Pian A, W. 
 
14 THE ATRIUM VEST^. 
 
 In a room at the east end of the house,' a small part of a more elaborate 
 painting was found,^ which suggests in general style the well-known fresco 
 of the house of Livia at Prima Porta. The pavements, which have been 
 preserved in several of the rooms, do not differ in style from those of the 
 Domus Vestalium described above.' 
 
 Construction and Materials: The walls of the republican period are 40 
 and 60 centimeters thick. Those of the earliest type are of opus quadratum 
 made from smaller blocks of the gray-green cappellaccio* which is found in 
 other early structures in the Forum and on the Palatine. These walls have 
 been replaced or added to by others composed of much larger blocks of 
 light-yellow and reddish-brown tufa. The walls of the later restorations are 
 of concrete faced with opus reticulatum or with brick, and vary in thickness 
 from 40 to 75 centimeters. The brick-faced walls, a considerable number of 
 which are still to be found, are of two types. In one of these the courses 
 of brick are from 3.50 to 4 centimeters wide and the layers of mortar from 
 1.50 to 1.75 centimeters. No bonding-courses are used. The bricks, which 
 are made wholly from flanged roof-tiles cut or broken in an irregular manner, 
 are magenta-red in color and of very fine texture. Walls of this type' are 
 found also in the Praetorian Camp and in the earlier buildings below the 
 temple of Augustus, as well as in the so-called Flavian rostra.' In the walls 
 of the other type, of which but a few remain, the courses of brick are but 
 3 centimeters or less thick and the layers of mortar 1. 25 to 1.75 centimeters. 
 The bricks are yellow or reddish-yellow in color and of remarkably fine 
 texture. Walls of this type, which are especially frequent in the facing 
 of tombs, where the brickwork was not intended to be covered by stucco 
 or marble, are not confined to any one period, though they are usually held 
 to have been most common during the reign of Nero or a little earlier. 
 
 'Plan A, XII. 
 
 2 This fresco, though faded, is still visible. 
 'Pp. 12-13. 
 * See p. 12, 11. 4. 
 
 ' The exact date of the walls of this type is yet unsettled. The greater number known to me 
 are of the time of Tiberius. Shortly after this period, they give place to an entirely new type. 
 ' This rostra is, as Huelsen holds, of the period of Augustus. 
 
III. 
 
 THE IMPERIAL ATRIUM OF THE FIRST PERIOD. 
 
 General Description: The new Atrium which arose after the fire of Nero/ 
 which we have called the imperial Atrium of the first period, was, as has 
 been said,^ wholly distinct from the earlier republican Atrium Vestae, as 
 it was also from the other imperial buildings inside the precinct.' It differed 
 from the republican Atrium in materials and methods of construction, as 
 well as in extent, orientation, and level. From the buildings of the period 
 also, it differed in orientation and level. In its size the new Atrium, which 
 was not out of keeping with the other buildings which arose out of the ashes 
 of the great fire, was a striking contrast to that of the Republic. Its length 
 was more than 70 meters, exclusive of the garden, which extended at least 
 15 or 20 meters further to the east. Its width was not less than 45 meters. 
 Beyond the Atrium on the north, there was also, opening upon a narrow 
 street, a line of small shops 4.50 meters deep, which were a part of the 
 building structurally and belonged probably to the Vestals. The street itself, 
 which connects the Forum directly with the height of the Velia, is to be 
 assigned to this period, since in this part of its course, it is built upon the 
 remains of the Domus publico, as are the shops opening upon it. The new 
 Atrium extended from the rear wall of the shops to the Nova Via on the 
 south.^ It was bounded on the west by the precinct of Juturna. On the 
 east its exact extent is unknown. It is probable, however, that on this side 
 it extended to the row of shops' opening upon the street in the rear of the 
 Atrium. The ancient altar" northeast of the Atrium was destroyed in the 
 course of the erection of these shops.' The new building was made to con- 
 form in its orientation to the south side of the Forum and to the new Nova 
 Via,* though the Regia and the temple remained unchanged. It is difficult 
 to fix accurately the level of this period. It can not have been, however, less 
 
 ' Plan B, cf. plan A. » Plan A, e, and p. 19. 
 
 > See above, p. 4. ' Plan A, m'-m". 
 
 ' The later Atrium and the temple of Vesta were " Jordan (/. c, 28) suggested that the change was 
 
 surrounded by a common precinct wall. brought about by the great mass of the 
 
 They were, however, structurally distinct temple of Venus and Rome. This is of 
 
 from each other. course impossible, since, as we now know, 
 
 ' The earlier Nma Via was at this time destroyed the first imperial Atrium was built a half- 
 
 and the present street built in its place. century earlier than the temple. The orien- 
 
 * Flan A, m^-rrl* , Towards the east also there may tation of the Domus Aurea, however, which 
 
 have been a few shops opening upon the differed but slightly from that of the later 
 
 Nova Via. Concerning the existence of temple (see Huelsen-Carter, The Roman 
 
 these shops, however, with the exception Forum, plate z) may not have been without 
 
 of one, the evidence is not conclusive. its influence. 
 
 15 
 
16 THE ATRIUM VEST^. 
 
 than one-half to two-thirds of a meter above that of the older building, the 
 walls of which have been left standing to that height.^ The center of the 
 court must have been somewhat higher, since almost a meter above the 
 earlier level a small piece of the natural rock, the cappellaccio of the neigh- 
 boring Palatine, has been left undisturbed.^ 
 
 Of the outer walls of the Atrium proper that on the north, which divided 
 the shops' from the residence rooms,^ is still standing to a considerable height, 
 except at the east end, where its presence is clearly indicated, however, by 
 the remains of the division-walls between the shops. The front wall of the 
 shops, as well as the corresponding wall on the opposite side of the street, 
 has been torn down.^ The wall dividing the Atrium from the temple area 
 also is in part traceable, though its course towards the west has been rendered 
 uncertain by later reconstructions." It may possibly have followed the line 
 of the later outer wall,' though no traces of it remain. Since, however, that 
 portion which can still be identified is in line with the later column wall, 
 it is more probable that the direction of its course did not in general differ 
 from that of the latter. On the south a portion of the outer wall behind 
 the rooms which remain is still standing several meters above the ground. 
 Beyond the Tablinum' towards the west, however, it has been wholly re- 
 built, though its earlier course is clear. The position of the wall bounding 
 the Atrium on the east is, as has been said, not definitely fixed. On the west, 
 between the precincts of Vesta and Juturna, the arches of the republican 
 period" which support the ramp to the Palatine are in part still to be traced. 
 The wall in their rear, which runs parallel to them, can not, therefore, at 
 any time have varied much in position. The foundations of the earlier wall, 
 moreover, are in certain places still visible beneath the restorations of the 
 later periods. The walls^" of the room between the court and the garden on 
 the east have been almost wholly destroyed. Their general position is, how- 
 ever, certain. 
 
 Plan and Arrangement of the Interior: The building as a whole is, so 
 far as can be at present determined, composed of two architectural units, 
 one of which is represented by the group of rooms on the north of the central 
 court" and the other by that on the south of it.'^ These two groups probably 
 were not distinct, when built, but were structurally united by the rooms 
 
 ' No traces of the pavement of this period renaain. ' See plan B. 
 
 2 The lucus Vesta covered originally the site of the ^ Plan A, 13. 
 
 Atrium court. It is possible that a small " For description of these arches, see Boni, Not. d. 
 
 portion of this grove, including the lotus Scavi, 1901, 62ff. 
 
 capillata, was left in the center of the court. '"Plate iv, fig. i . (From a photograph taken ur- 
 'Plan A, m-m. ing the course of the excavations in 1903, 
 
 * Plan A, 1-9. by the courtesy of Director Boni.) The 
 
 ' The foundations of the front wall of the shops walls are seen at the level of the foundations 
 
 remain throughout their whole length. at the right side of the figure. Concerning 
 
 Concerning the massive concrete founda- the period of this room, see p. 18, n. 8. 
 
 tions of these two walls, see p. 18, u. 10. "Plan A, 1-9. 
 
 Plan E, 56a-56c. "Plan A, 10-15. 
 
E. a. V D. 
 
 Fig, 1. Foundation of the Walls of the Imperial Atrium of the First Period. 
 
 Fig. 2. The Republican Altar. 
 
THE IMPERIAL ATRIUM OF THE FIRST PERIOD. 17 
 
 on the west, which have been destroyed in the course of later restorations. 
 The two parts formed thus a single complex structure, which is plainly of 
 one period and the work of one builder. This is shown most clearly by the 
 general harmony in architectural plan and by the structural unity which 
 exists throughout the parts of the building which are preserved. The level of 
 the concrete foundations also varies but a little in the several parts, although 
 that of the pavement above them can not be determined. The methods of 
 construction are, in general, the same and the character of the material used 
 is identical. 
 
 The interior of the Atrium can be reconstructed only in its more general 
 features. The size and arrangement of the rooms on the west it is impossible 
 to determine, since they have been rendered inaccessible by the later build- 
 ings which have been placed on top of them. The existence and general 
 position of the rooms on the north^ is certain, but no exact restoration of them 
 is at present possible. While, however, the reconstruction of the Atrium as a 
 whole is, as has been said, impossible, the general plan of the building is 
 clear.^ In the center was a large court,' or atrium, extending from north- 
 east to southwest, 45.50+ meters long and about 21 meters wide. On both 
 sides and possibly at the west end of this court was a series of lofty rooms.* 
 There was left on the east a broad entrance into the garden beyond,' which 
 sloped gradually upward to the rear of the Atrium. The main entrance" 
 into the building was through a vestibule directly from the street on the 
 north and not, as at a later time, from the temple area.' The door on the 
 north, which was closed by a cross-wall of the house of the next period, was 
 1.75 meters wide and between 2 and 3 meters high. There was doubtless, as 
 later, a door leading directly from the Atrium into the temple area for the use 
 of the priestesses in their attendance on the temple. The rooms on the north 
 were at least nine in number. They were 1 1 meters or more long and vary in 
 width from 4.15 to 4.85 meters. Between the three rooms towards the west, 
 there are doors 1.77 meters wide.' There may have been originally similar 
 doors between the other rooms also, though no trace of them now remains, 
 since the walls have been destroyed almost to the foundations. For the same 
 reason it is not possible to decide whether the rooms opened into the court 
 through doors or wide arches. On the south side of the court opposite the 
 entrance was a large open room,^" which on account of its position and pro- 
 portions may be regarded as the tahlinum. Next to it on either side was a 
 smaller room, which was open also to the atrium. Of these rooms that on 
 
 ' Plan A, 1-9. ' Plan A, 1-9, 10-15. 
 
 2 See plan A. * Plan A, c. 
 
 ' From data recently obtained it seems possible that ' Plan A, n. 
 
 earlier walls existed in the same position as ^ See plan B and p. 23. 
 
 those on which the ends of the colonnade * Plan A, 1-3. 
 
 of the next period (see plan B) rested. ^Approximately 6 Roman feet, a common meas- 
 
 The earlier court would, in that case, have urement throughout the whole building. 
 
 extended on the east and west only to these ^''Plan A, 13. This room was 16.07 meters wide 
 
 walls. and 10.28 meters long. 
 
18 THE ATRIUM VEST^. 
 
 the east^ was 7.43 meters wide, while the corresponding room,^ though, in 
 all probability, of the same width originally, was at a later time 15 cen- 
 timeters narrower. The latter room, which at least later contained the shrine 
 of the household gods,' may be held to be the lararium of the house. It is 
 probable, therefore, that the two rooms were throughout the early periods 
 of the Atrium's history regarded as the alie.^ Their position is, however, 
 unusual. Beyond the ala on the east is a smaller room,' which is entered 
 from the court by a door 2.50 meters wide. On either side of this room 
 there were doors 3 meters wide and 3.70 meters high. Still further to the 
 east was another room" 5.10 meters long, which was entered from a narrow 
 corridor or directly from the garden. Whether there were other rooms 
 beyond this, opening upon the garden, it is not possible to determine. At 
 the east end of the court, extending into the garden, was a single room,' 
 which resembles in its position the tablinum of the Graeco-Roman house. 
 It is possible that this room' was open towards both the atrium and the 
 garden, though no proof exists, since the walls have been in large part 
 destroyed. The garden, which was a part of the earlier lucus Vesta, occupied 
 the remaining part of the Atrium on the east. Concerning the rooms on the 
 west no data are now obtainable. The existence of a wall on the west of the 
 court may, however, be assumed as certain. The position suggested in 
 the plan is that of the later wall, which is the same distance, within a few 
 centimeters, from the tablinum as the corresponding wall on the east of 
 the court. Next to the temple there were doubtless, as later, rooms for the 
 use of the cult. 
 
 With the exception of the tablinum and ala on the south, the purpose of 
 the various rooms is not clear. The rooms adjoining the entrance on the 
 north were, probably, more public in their character. It is probable that 
 the kitchen and the rooms connected with it were in the more remote part 
 of the house east of the tablinum. The sleeping rooms were doubtless, as 
 later, in the upper story, if there was one. 
 
 The small rooms, or shops, on the north,' which are a part structurally 
 of the building, are eleven at least in number, exclusive of the one'" which 
 was used as an entrance vestibule to the Atrium. They are 4.50 meters 
 long and correspond in width to the adjoining rooms in the Atrium. The 
 front wall of these shops, as well as the corresponding wall on the opposite 
 side of the street, has been, as stated above, destroyed to its foundations." 
 Of the travertine posts in front of the shops portions of but two remain. 
 
 ^ Plan A, 12. * It is possible that this room and the walls adjoin- 
 
 2 Plan A, 14. ing it on either side are to be assigned to 
 
 ' See plan D, 42, and p. 33. the next period. Cf. p. 17, u. 3. 
 
 * There are no clearly defined a/« elsewhere. The • Plan A, m-m. 
 
 rooms suggested must, therefore, be ac- '"Plan A, room adjoining 3. 
 
 cepted as such. *^A full discussion of these concrete foundations, 
 
 ' Plan A, II. This room is 5.50 meter? wide. which rise in certain parts more than a 
 
 • Plan A, 10. meter above the republican level, will be 
 ' Plan A, 1). presented by the writer at a later time. 
 
THE IMPERIAL ATRIUM OF THE FIRST PERIOD. 19 
 
 With only a few exceptions, however, they can be restored from the impression 
 left in the concrete of the later pillars, which were built in front of them. 
 
 The Republican Altar: The ancient altar,' the remains of which are still 
 to be seen in the room at the northeast corner of the Atrium, was at this 
 time destroyed. This altar is made of ashes and sacrificial material and was 
 surrounded by a narrow gutter. At a very early period it stood probably 
 inside a separate precinct, or templum, which was inclosed by a wall. At a 
 somewhat later period, possibly at the time of the abandonment of the altar 
 as a place of sacrifice, a second wall of opus quadratun^ was built inside the 
 precinct wall for the better protection of the altar itself. Of this inner wall 
 and the altar inclosed by it but little now remains.' 
 
 Architectural Details and Construction: The height of the rooms can not 
 be ascertained. That they were lofty may be assumed from the thickness of 
 the walls, and the height of the doors which remain. There are no pavements 
 left in any of the rooms, and there is no evidence of the use of hypocausts, 
 as in the later Atria. All traces of decoration have disappeared, with the 
 exception of a small bit of fresco in one room. The large quantity of rare 
 marbles, however, which have been broken in pieces for use as filling in the 
 concrete foundations of the next period,* points to their extensive use as 
 decoration in the earlier building. There are no stairs and no traces remain- 
 ing of an upper story, though it is probable that one existed.^ There is 
 beneath the whole Atrium an intricate network of sewers." No attempt has 
 been made to describe these, since the data available are insufiicient. 
 
 In methods of construction as well as in the materials used, the Atrium 
 is perfectly in agreement with the other buildings of the period of Nero. 
 The walls are throughout of concrete faced with brick.' The outer walls 
 are 89 centimeters' thick. The inner walls on the south, where, because of 
 the width of the rooms, the vaulting supported by them was heavier, are of 
 the same thickness. The inner walls on the north are but 74 centimeters, ° 
 corresponding to those of the shops in their rear. Bonding-courses of tegulee 
 bipedales^" are not found in any of the walls on the north. They have been 
 
 ' Plan A, B; plate iv, fig. 2. The divinity to whom ' Professor Huelsen has suggested that walls of 
 
 the altar was dedicated is unknown. The such thickness would not have been built 
 
 shrine of Aius Locutius lay within the lucus had there been no upper story to support. 
 
 Vesta. Huelsen, from the description of its ' Under the careful direction of Commendatore 
 
 position given by Cicero ((feZ)ro., J, 45, 101) Boni, a number of the sewers have been 
 
 locates it however much further to the west restored to their original use. 
 
 along the Nova Via. The worship of the ' The use of the term brick-walls should be avoided, 
 
 divinity to whom the altar belonged was since none existed in Rome, 
 
 probably continued, since in the wall be- ^ This measurement, which is equal to 3 Roman 
 
 hind there is found a niche for a statue. feet, Is very common in walls of this 
 
 • The inner wall is of a much later period than the period and later. 
 
 altar itself. "Another common measurement, which is equal to 
 
 ' See plan A, «. 2.5 Roman feet. 
 
 * See p. 28. "See p. 3, n. 4. 
 
20 THE ATRIUM VEST^. 
 
 used occasionally, however, in certain of the walls on the opposite side. 
 The upper portion of these walls is, however, in all probability, a restoration 
 of the succeeding period. With the rise of the new city a distinct change 
 takes place in the type of brick-facing used for concrete structures. The 
 courses of brick, which are 3.75 to 4.25 centimeters wide, show much less 
 variation in the individual walls. The layers of mortar also, which are 1.25 
 to 2 centimeters thick, become more regular. The change, or rather develop- 
 ment, is especially noticeable in the character of the bricks themselves. 
 The use of broken tiles for bricks, which was the most marked characteristic 
 of the preceding period, was abandoned,^ and the bricks assumed the more 
 regular triangular shape. The new bricks are of a much coarser quality 
 than the tile-bricks and are much looser in texture, though they are well 
 burned. In color they vary from yellow to yellow-red. The concrete, also, 
 is of a coarser composition, though otherwise of a good quality. The filling 
 is in large part of tufa and travertine with but a small proportion of broken 
 brick and marble. Both mortar and concrete are marked by their unusual 
 gray tone, which arises from the almost entire absence in their composition 
 of red pozzolana? 
 
 ^ The walls in which the bonding-courses appear ' In the description of the modes of construction 
 
 most frequently are those east of the tabli- and materials used in the different periods, 
 
 num. Bonding-courses appear sporadically all minor details have been omitted except 
 
 in this period, but their use is not common such as are necessary for the differentiation 
 
 until the time of the Flavians. between the various periods. The whole 
 
 2 At several later periods they are again used in subject will be treated fully by the writer 
 
 considerable quantities. at a later time. 
 
IV. 
 
 THE IMPERIAL ATRIUM OF THE SECOND PERIOD. 
 
 General Description: A short time only after the completion of the new 
 Atrium by Nero, a considerable part of it was destroyed by fire. In con- 
 sequence of this partial destruction many of the walls of the building were 
 rebuilt from the foundations and extensive changes were made in its general 
 plan.' From the walls which remain it is evident that the western part of 
 the structure suffered most severely. The exact period of the restoration of 
 the walls, on the south side of the building, east of the tablinuni' is uncer- 
 tain. The wall west of the tablinum was, however, at this time rebuilt from 
 about 3 meters above the ground.' On the north side the wall next to the 
 court* and the division walls between the rooms,^ with the exception of 
 one, were destroyed to the level of the later pavement. The back wall 
 of the rooms," on the contrary, and the walls of the shops beyond' were 
 left untouched, at least to a considerable distance above the ground. It 
 is possible, therefore, that the destruction of the other walls' on this side 
 was partly due to the changes in the general plan of the building, which 
 took place at this time. The rooms on the northeast beyond the court" 
 were not, so far as can now be ascertained, affected by these changes. It 
 is probable, therefore, that the fire did not extend so far in this direction. 
 Many pieces of marble showing traces of fire were found, in the course of 
 the excavations, in and near the pozzi at this end of the court. These may 
 have been, however, from some more remote part of the building or have 
 been injured in some later fire.'" The rooms on the west" were, it is probable, 
 wholly destroyed, since the walls are rebuilt from the foundations." Whether 
 the foundations themselves belong to this or to the preceding period can not 
 be determined, except in a few cases, since they are for the most part con- 
 cealed by later structures. The changes in plan on the north side of the 
 building suggest, however, a corresponding change in the arrangement on 
 this side also. New foundations would in that case have been required. 
 
 iPlanB. C/. plan A. "Plan 6,8-9. 
 
 'Plan 8,13. "Seep. 43. 
 
 'The line is plainly marked by the change in "Plan B, 15-28. The arch in which was found a 
 
 brickwork. brick-stamp of the second century mentioned 
 
 * Cf. plans A and B. by Huelsen (Roem. Mitth., 1889, p. 246, n.) 
 " plan A, 1-7. may very well have belonged to a later restor- 
 
 • C/. plans A and B. ation. 
 
 ' plan B, m—m. ''The outer wall on the west, though in large part 
 
 ' The wall next to the court and the division walls restored at this time, was not wholly de- 
 
 between the rooms. stroyed. 
 
 21 
 
22 THE ATRIUM VEST-ffi. 
 
 In size and extent the Atrium remained unchanged,' except on the side 
 towards the temple, where a small portion of the sacred area was included in 
 the court. The level was raised 30 to 50 centimeters above that of the pre- 
 ceding period, that is 97 to 100 centimeters above that of the republican 
 building.^ The rooms on the southwest' may have been, as later, 30 centi- 
 meters higher than this. The new rooms on the north* were on the same 
 level as the outer corridor upon which they opened.^ The rooms beyond 
 the Atrium on the west," which had no direct connection with the other parts 
 of the building, were at least 30 centimeters below its level, being but 70 
 centimeters above that of the republican period. 
 
 Plan and Arrangement of the Interior: The differentiation of the new 
 walls of the Atrium from those of the preceding period is not difficult, since 
 they are not only dissimilar in their mode of construction, but are structurally 
 independent. The determination of the relation which these new walls bear 
 to each other is no less easy. That they are parts of a single structure is 
 evident from their agreement in methods of construction as well as from 
 their conformity to a general architectural plan. The unity which exists 
 among them is, however, most clearly shown by the structural continuity 
 by which they are marked.' The general level is consistent and that of the 
 foundations is uniform throughout. 
 
 In general plan the Atrium of the second period was a development 
 rather than a mere restoration of that of Nero. The central court, or atrium, 
 did not probably differ in length from that of the preceding period; its 
 width was, however, somewhat increased by the pushing back of the wall 
 on the north. Since the new wall varied slightly in direction from the older 
 one, a difference arose in the width of the court at the two ends, which 
 became much more marked after the extension of the court to the east.' 
 On all sides of the new court there was added a colonnade, or portico, of 
 about 4 meters in width." Auer,'" who maintained that the colonnade 
 consisted of but one story, in order to reconcile the height" of the columns, 
 which can not have exceeded 5 or 6 meters, with that of the windows, held 
 that the portico was covered by a slanting roof, which rested upon the 
 entablature above the columns and was attached to the opposite wall below 
 
 ^C/. plans A and B. ^ This structural continuity is most noticeable on 
 
 ' The level of the colonnade is taken as the standard the southwest. It exists, however, through- 
 
 for the period. See plate ii, fig. i; wall on out. 
 
 the right. ' See p. 36. 
 
 8 Plan B, 14-23. "The colonnade on the north and east is 3.60 
 
 *Plan B, 2-7. meters wide and on the west 3.90 meters; 
 
 ' If the hypocausts which remain in several of the on the south it varies from 4 to 4.15 meters, 
 
 rooms are of a later period, the level of the The foundation wall of the colonnade at 
 
 rooms was a half-meter lower than that of the east end of the court is seen below the 
 
 the corridor. It is more probable, however, foundation of the octagonal structure on the 
 
 that the hypocausts were built at the same left in plate iv, fig. i. 
 
 time as the rooms. '"Auer, Der Tempel der Vesta, 4. 
 
 « Plan B, 24-28. " See p. 23, n. 2. 
 
THE IMPERIAL ATRIUM OF THE SECOND PERIOD. 23 
 
 the windows of the second story. It is more probable that the colonnade 
 was composed of two rows of columns/ one above the other, the lower one 
 of which was from 5 to 6 meters in height' and the upper somewhat less. 
 There was no upper floor corresponding to the intermediate entablature,' 
 since the columns of the lower arcade were not sufiiciently high to carry 
 it to a level above the windows of the lower story. The floor would, more- 
 over, have cut off the light almost wholly from the inner corridor. There 
 were in each story thirty-two columns.^ The travertine foundations on 
 which the bases of these rested have been in large part preserved. The 
 intercolumnar spaces vary slightly, increasing towards the east of the atrium 
 from 3.30 to 3.43 meters. The variation is, however, not regular.' In the 
 center of the court was a basin* into which steps descended from either end. 
 The water contained in this basin was for the ordinary uses of the house- 
 hold.' That designed for use in the rites connected with the temple worship 
 was kept in the rooms set aside for the cult.' The entrance into the Atrium 
 was from the area in front of the temple. The door which had in the pre- 
 ceding period led directly from the street on the north was closed by one 
 of the newer walls." Between the posts of the new door and the columns 
 opposite there was, on either side, a single column, which supported possibly 
 a vaulted vestibule. Beyond the main entrance was a smaller door which 
 led directly into the sacred precinct of the temple." This was closed later 
 by the erection of the cedicula}^ 
 
 The rooms on the south belonging to the preceding period'^ were not 
 altered at this time, except by the addition possibly of the small posts in 
 front of the alce^ by which these rooms were separated a trifle more from 
 the corridor outside. The remaining rooms, on the north and west, may be 
 divided into two groups.** The group on the north, which replaced the 
 earlier rooms opening directly upon the court, consisted mainly of two large 
 rooms" united by a narrow corridor," from which opened three or possibly 
 
 ' Jordan, Der Tempel der Vesta, 37. 'Plan B, o. The basin was 14 meters long, 4 
 
 2 The height of the columns has been estimated meters wide and 5 deep. The upper part 
 
 from the size of the travertine blocks on has been restored. 
 
 which their bases rested. The marble bases ' No water could be used for the sacred rites except 
 
 and columns which have been found belong such as had been brought fresh from foun- 
 
 to a much later period. They do not, how- tains or running streams in vessels specially 
 
 ever, differ materially, in all probability, prescribed for the purpose. See p. 17. The 
 
 from those of the earlier colonnade. basin, on the other hand, was supplied with 
 
 ' There was a similar col onnade in the building of water drawn from ordinary sources and 
 
 Eumachia at Pompeii (Mau-Kelsey, Pom- conducted into the Atrium through lead 
 
 peii, 113-114). pipes. 
 
 * The number of columns at the sides of the colon- 8 plan B, 24-27. 
 
 nade is twelve, that at the ends, six. Jordan 'PlanB. The wall between 2 and 4. 
 
 (/. c.) holds, though in my opinion without I'See plan B. This inclosure was not entered by 
 
 sufficient reason, that the occurrence of the the public. 
 
 number six is not without significance. "Flan C, E. 
 
 ' Schulze (Jordan, /. c.) considers the variance due "Plan B, 10-15. 
 
 to regard for perspective. The size of the *'Plan B, 12 and 14. 
 
 earlier Atrium at least is not such as to "Plan B, 1-7; 15-28. 
 
 warrant this conclusion. See pp. 43-44 for "Plan B, z and 7. 
 
 further discussion. "Plan B, 3. 
 
24 THE ATRIUM VEST^. 
 
 four smaller rooms. ^ The walls of the newer rooms were not at right angles 
 to the older wall in their rear, to which they were attached; this caused a 
 slight irregularity in the shape of the larger rooms at either end. The first 
 room on the north^ remained unchanged, except that its length was some- 
 what lessened by the change in position of the front wall. The room adjoin- 
 ing it,' which is 8.30 meters long, was on the contrary widened to 7.30 
 meters. The width of the corresponding room on the east can not now be 
 determined. The corridor by which these rooms were united is 3 meters 
 in width. The small rooms* which open upon it are only 4.59 meters 
 long. The doors leading from the corridor into the smaller rooms, as 
 well as those by which the rooms were united to the colonnade outside, 
 can not be restored, since the walls are destroyed almost to the founda- 
 tions. No change took place in the rooms situated beyond the court and 
 opening upon the garden.' The group of rooms on the west and south- 
 west' is much larger than that just described. On the south, next to the ala 
 on the west of the tahlinum, is a room" 4.14 meters wide, which is to be con- 
 nected rather with the older rooms adjoining it on the east than with the 
 newer ones on the west. Beyond this room, at the southwest corner of the 
 Atrium, are a number of rooms, which, though connected structurally with 
 the rest of the building, seem from their position and arrangement to form 
 a distinct group by themselves. The center of this group was a large room,* 
 at the west end of which was an apse. On the south were three smaller 
 rooms, ° a little higher in level than the larger one and connected with it 
 by wide arches. Next to these rooms on the north is a stairway,'" which gave 
 direct access to them from the clivus, or ramp, leading to the Palatine. 
 Beyond the stairs is a narrow corridor," resembling that on the north, out of 
 which opened a room*^ 5.30 meters long and 4.71 wide. At the northern 
 end of the corridor is a larger room'' which may have opened directly upon 
 the colonnade." It is uncertain whether the rooms just mentioned formed 
 a part of the group of which the large room was the center or were con- 
 nected with the Atrium proper. On the west of these rooms and wholly 
 distinct from them is a long hall,'' divided into three closely connected rooms 
 by low archways." At the further end of this hall is a vaulted storeroom," 
 
 1 plan B, 4-6 and plate v, fig. 2. * Plan B, 16-18. The rooms were 4 meters long 
 
 2 Plan B, I. and over 3.50 meters wide. 
 ' Plan B, i. '"Plan B, 20. 
 
 <Plan B, 4-6. These rooms are 5.13, 2.95 and "Plan B, 21. 
 
 3.38 meters wide. ^VUn B, 22. 
 
 I" Plan B, 8-9. The wall in front of the new rooms "Plan 6,23. This room is 8.90 by 7.40 meters in 
 
 is not continued beyond the division wall size. 
 
 between rooms 7 and 8, which is of the pre- "There are remains of a window opening upon the 
 
 ceding period. colonnade at a later period. 
 
 « Plan B, 15-28. "Plan B, 25-27, and plate v, fig. i. 
 
 ; Plan B, 15. '"The walls by which the arches have been filled 
 ' Plan B, 19. This room, which was divided into in may be of the same period; it is more 
 
 two unequal parts, was 12.64 ™eters long probable, however, that they were built a 
 
 and 6.00 meters wide. little later. 
 
 "Plan B, 24. 
 
PLATE V. 
 
 E. 8. V. D. 
 
 Fig. 1. The Rooms for the Use of the Cult. 
 
 E B V. D 
 
 Fig. 2. I HE North Side of the Atrium. 
 
THE IMPERIAL ATRIUM OF THE SECOND PERIOD. 25 
 
 in which were found many broken amphorae and other vessels.* The only 
 entrance to these rooms was from the temple precinct.^ 
 
 No change took place in the shops on the north, except that by the closing 
 of the door into the Atrium' one was added to their number. It is possible 
 that the shops along the Nova Via* were added at this time, but the scanty 
 remains belong to a later period. 
 
 The Sacellum Larum: The rooms on the southwest,^ though structurally a 
 part of the building, formed, as has been said, a distinct group by them- 
 selves. The stairway on the west, by which the Atrium was connected 
 with the ramp leading to the Palatine, though affording admission to the 
 other parts of the building, was designed especially to give direct access to 
 these rooms. The communication with the outer world, especially with the 
 Palatine, was at a later time made still more easy by the erection of a second 
 stairway," by which the court was connected directly with the Nova Via. At 
 the further end of the court was a large apse. In the wall at the end of 
 the court toward the east, by the building of which the rooms were still 
 further cut off from the Atrium, there were added also three niches for 
 statues.'' The group of rooms must, therefore, have had a religious purpose' 
 and have been designed for some cult over which the Vestals had special, 
 though not exclusive, oversight. The remoteness of the rooms from the 
 temple and their lack of any direct communication with it, as well as the 
 freedom of intercourse with the outer world indicated by the stairways, for- 
 bid the identification of this cult with that of Vesta and the Penates, the 
 center of which was the temple. A close connection, however, must have 
 existed between them. Of the various cults connected with, though distinct 
 from, that of Vesta and the Penates, the most nearly allied is that of the 
 Lares Publici. Their shrine, which was known as the Sacellum Larum, 
 in distinction from the tsdes Larum, was recognized by the Romans as one 
 of the determinative points in the line of the Pomerium of the Palatine city." 
 It must have been situated, therefore, within a short distance of this corner 
 of the Atrium, since in this vicinity the line of the Pomerium, which on the 
 north followed in general the course of the later Nova Via, turns abruptly 
 toward the west. It has seemed to me probable, therefore, that in this group 
 of rooms, under the same roof with, though distinct from, the shrine of the 
 other gods of the state hearth, is to be found the hitherto unplaced Sacellum 
 Larum. 
 
 ' Bull, comun., xxxi, yof. Huelsen-Carter, Roman * The so-called libation-bowl in the pavement of the 
 Forum, 2i6f . court had no connection with religious rites, 
 
 2 The door connecting these rooms with the Atrium for which it was wholly inappropriate on 
 
 is not original. account of both its size and construction. 
 
 3 Plan A, rj. It was probably, as Professor Mau has 
 
 4 Plan C, m"'. suggested, connected with the cleaning of 
 'PlanB, 16-19. the pavement, since it leads directly into 
 e Plan E, 44. the sewer below. 
 
 ' See plan E, 47. ° Tac, Ann., xn, 24. C/. Richter, Top., plate 2. 
 
26 THE ATRIUM VESTiE. 
 
 The Cult Rooms: The presence in the rooms on the west' of certain 
 structures held to be ovens,^ as well as of a number of amphorae and frag- 
 ments of other vessels,' has led to the almost universal acceptance of these 
 rooms as the private kitchen and storerooms of the Atrium. That they 
 were not intended for such a purpose is, however, clear from the primitive 
 character of the vessels found and the noticeable absence of any of the more 
 common appliances for cooking,^ as well as from the existence of a regularly 
 appointed kitchen' at the opposite end of the Atrium. It is evident, more- 
 over, from the complete isolation of the rooms and from their lack of any 
 means of communication" with the other parts of the building, that they 
 were not even a part of the Atrium but were intentionally separated from 
 it. Their proximity to and direct communication with the temple precinct, 
 on the other hand, make it certain that they were instead connected with 
 the temple and belonged to it. From the unusual arrangement of the rooms 
 themselves and the peculiar character of their contents' it is evident that they 
 were designed for the special use of the cult and for the storing of the instru- 
 mentum* belonging to it. In addition to the instrumentum there were no 
 doubt preserved in these rooms such of the pignora imperii^ as could not 
 conveniently be kept in the temple itself. It is probable that the documents 
 of state and the other articles of value which were entrusted to the Vestals for 
 safekeeping were kept here rather than in the private rooms of the Atrium. 
 Of the objects found in the rooms the pointed amphorae, which are of the 
 primitive type prescribed for the use of the cult,'" were for the carrying of 
 the water used in the sacred rites and in the cleansing of the temple. The 
 smaller vessels, of which numerous fragments have been found, were for 
 the use of the Vestals in the preparation of the sacred cakes and of the other 
 materials for sacrifice and purification over which they had charge." In 
 one of these vessels there are still preserved the remains of pastry.'^ The 
 appliances for the grinding of the sacred meaP' and for the preparation of 
 the muries^^ were also of the simplest kind and have been either carried 
 away or destroyed. In one of the rooms" is a basin of peculiar construction. 
 It consists of a square structure 3.1 1 meters long and a meter wide, which, 
 
 * Plan !B, 24-27. ' For the instrumentum of the temple, see Wissowa, 
 
 * See p. 44. Religion und Kultus, 4o6f. 
 
 ' A number of these are still preserved in the store- " The size and arrangement of the temple make it 
 room in which they were found. improbable that the less important sacra 
 
 * There are no traces of stoves or of any of the were kept there. 
 
 ordinary receptacles for water. ^®A number of these are still to be seen in the room 
 ^ The kitchen now seen belongs to a later period. in which they were found (plan B, 24). 
 
 The earlier one must have been, however, "For the principal duties of the Vestals in con- 
 in the same part of the Atrium, where the nection with the cult, see Wissowa, /. c, 143. 
 
 more private life of the household was "Boni (Vaglieri, Scavi Recent! nel Foro Romano, 
 centered. 71, n. 1) recognizes in this the strues men- 
 
 "Th^ door leading from the Atrium into these tioned by Festus (p. 310, n.). 
 
 rooms is not original. "Wissowa, /. c. 
 
 ' Especially the rude pottery and the basin de- "Plan B, 27. 
 scribed below. 
 
THE IMPERIAL ATRIUM OF THE SECOND PERIOD. 27 
 
 though inclosed regularly at both ends by walls, is without either front or 
 back.' The bottom of the basin, which like the ends was lined with opus 
 signinum, slopes noticeably towards the back, where it drains directly into 
 the sewer below. This basin, or rather drainage sink, was probably for 
 the washing of the vessels used in the temple service, which were cleansed 
 by pouring water over them directly from the amphorae in which it had been 
 brought.^ By this method the water as well as the basin was kept free from 
 all impurities, which were carried immediately into the sewer behind. The 
 room' or rooms next to the temple area were probably for the servi publici 
 attached to the temple. 
 
 Architectural Details and Construction: The height of the various rooms 
 can not be determined. In the central court as well as in the corridor sur- 
 rounding it traces of the pavement of several periods still remain. The 
 earliest of these, which is of opus spicatum, is, however, not older than the 
 second century. No pavements have been preserved in the rooms on the 
 north or on the south. In the group of rooms on the west and southwest* 
 the larger room and the smaller ones adjoining it are paved with white and 
 black mosaic. The pavement in the smaller rooms is probably original, 
 since above it have been found remains of a pavement of opus sectile of the 
 style popular in the time of Hadrian.^ The pavement in the larger room 
 belongs probably to a later period.' In the rooms on the north' hypocausts, 
 or at least double floors,' were built. In several of the rooms these have been 
 preserved. Hypocausts of this period remain also in the smaller rooms on 
 the southwest. That in the larger room on the west was inserted at a later 
 time. In one of the rooms on the north remains of a stairway have been 
 found. It is probable, therefore, that a second story extended over the whole 
 house, with the exception possibly of the rooms on the southwest,' though no 
 traces of it are left. But little change took place, probably, in the purpose 
 to which the various rooms of the Atrium were set apart. The sleeping- 
 rooms were without doubt transferred to the upper story, if this had not 
 already been done in the preceding period. 
 
 As in the other buildings of the period of Domitian,*" the walls are clearly 
 distinguishable from those of the preceding period, both by their methods of 
 construction and by the materials used. The outer walls, especially those on 
 the southwest, while not free from restoration, have in no case been rebuilt 
 
 ^ On account of the lowering of the level of the ^ Traces of double floors remain in several of the 
 
 room by the removal of the ancient pave- newer rooms. In one of these there is also 
 
 ment, the basin at present resembles a low evidence of the presence of fire. Professor 
 
 platform. Mau holds that the use of such floors is not 
 
 ' See p. 23, n. 7. sufficient proof of the existence of hypocausts. 
 
 ' Plan B, 28. • The character of the rooms makes the existence 
 
 'Plan B, 16-23. of a second story improbable. 
 
 'See p. 33. >°0f the buildings of Domitian, those of especial 
 
 " The pavement is of a much coarser type than interest are the Domus Augustana and the 
 
 that in the adjoining rooms. so-called Stadium on the Palatine, the 
 
 ' Plan B, 2-7. temple of Augustus, and the villa at Albano. 
 
28 THE ATRIUM VEST^. 
 
 from the foundations. Their width, therefore, remains the same as in the 
 preceding period. The' wall on the north of the court also is at the west end 
 89 centimeters wide, though further towards the east, a few courses above the 
 foundation, it has been reduced to 60. The walls on the west are, with one 
 exception, 60 centimeters, although those on the north are 75, as were the 
 older ones. The walls in which are set the travertine bases for the columns 
 are 89 centimeters. Bonding-courses, which are found only sporadically, 
 if at all, in the earlier walls, appear regularly in those of this period. They 
 are especially noticeable in the rooms on the southwest, where the tegulcs 
 hipedales of which they are made are 4.50 to 5 centimeters thick and of a 
 dark-red color. Since the walls on the north have been almost wholly rebuilt, 
 only one bonding-course remains ; upon the shelP formed by this rests the 
 lower floor of the hypocausts. The courses of brick measure from 3.75 to 
 4.25 centimeters, as in the last period, and the layers of mortar from 1.50 
 to 2 centimeters. In the materials used, the type of construction differs from 
 that of the time of Nero. The bricks are of the same type, many of them 
 possibly left over from the earlier buildings. The mortar, however, is unlike 
 that of the preceding period, showing a large proportion of red pozzolana. 
 The concrete, especially that used in the foundations of the walls, is marked 
 also by the large quantities of rare marbles which have been used as filling.^ 
 
 1 This shelf, which is formed by the change in the width of the wall, at a little distance above the 
 
 foundations, from 89 to 75 centimeters, is characteristic of this group of rooms. 
 * This is best seen in the foundations of the column wall on the west and of the wall opposite. 
 
E. B. V. D, 1902, 
 
 Fig. 1. The Rooms on the South, built by Hadrian. 
 
 Fig. 2. Walls of the THiRD and Fourth Periods. 
 
THE IMPERIAL ATRIUM OF THE THIRD PERIOD. 
 
 There is no evidence that any further calamity befell the Atrium until the 
 time of Commodus. At two different periods before that time, however, 
 important changes took place in the building. The first of these, which 
 represents the third stage in its growth,' consisted in the addition by Hadrian 
 of the group of rooms on the east,^ the center of which is the large hall, or 
 exedra, and of a smaller group on the south.' These two groups will be 
 discussed separately. 
 
 The Rooms on the East: The group of rooms on the east, which was 
 held by Auer* to be the oldest part of the imperial Atrium, was located 
 at the rear of the garden and united to the earlier structure by the outer 
 walls only.^ The space occupied by it was at least 42 meters long and i6 + 
 meters wide. The new building extended on the north only to the line of 
 shops belonging to the first period." The destruction of these shops and the 
 inclusion of the space occupied by them in the Atrium took place at a later 
 time.' On the south the new building extended to within a short distance 
 of the Nova Via.* The rooms extended on the east to the street in the rear 
 of the Atrium, ° the shops opening upon which were in consequence almost 
 wholly destroyed.'" The orientation of the new rooms conforms in general 
 to that of the earlier building, but their level differs greatly, being a meter 
 higher than that of the rooms on the west, that is to say, two meters above 
 the republican level. Since the height of the hypocaust openings, which are 
 still to be seen in the smaller courts" on the north and south, is but 10 centi- 
 meters less than this, the original level of the rooms can not have been lower 
 than it is at present. This noticeable difference in level between the newer 
 and the older parts of the building was due to the natural rise in the ground 
 towards the east and south. The garden, at least the eastern end of it, agreed 
 in level with the new rooms which adjoined it. This is evident from the 
 
 ^ See p. 6. ' The walls at this corner of the Atrium are of the 
 
 2 Plan C, 29-39. °^^^ period. 
 
 ' Plan C, 13 a-dj and plate vi, fig. I. ^ The rooms on the south of the court at a higher 
 
 * Auer, Der Tern pel der Vesta, 20. level are of a later period. 
 
 ^ There may have been a row of rooms along the ^ Walls 1 .20 to 1 .30 meters in thickness were built 
 
 south side of the garden, but no traces of behind the new rooms, to resist the pressure 
 
 them remain. of the earth by which the street had been 
 
 ' The point at which the new front wall was raised to the higher level. 
 
 attached to the rear wall of the shops may "The walls between the shops were in no case 
 
 still be determined from changes in the wholly destroyed. 
 
 masonry. "Plan C, 31 and 39. 
 
 29 
 
30 THE ATRIUM VESTj^I. 
 
 height of the concrete foundations of the new court on the north, which 
 projected 6o centimeters beyond the front wall.' These foundations, which 
 are but a few centimeters lower than the pavement of the new rooms, must 
 have been, at the time they were built, below the level of the garden and 
 concealed by it.^ Moreover, in front of the exedra traces remain of a sewer, 
 the top of which is but a little lower than the pavement of the rooms under 
 which it ran. This sewer, traces of which may still be seen on the right of 
 the steps leading to the exedra, must also have been originally below the 
 level of the garden. It is clear, therefore, that the garden was either raised 
 consistently to the level of the new rooms or, as seems more probable, made 
 to slope gradually up to them. The walls of the new building, with a few 
 exceptions, are still standing to a considerable height, though they have 
 suffered much from restoration in the next period.' 
 
 Arrangement and Description of the Rooms: Since the time of Auer* these 
 rooms have been generally recognized' as belonging to one period and as 
 forming a distinct group by themselves." That this conclusion is correct 
 is very evident from their harmony in architectural plan and their uniformity 
 in methods of construction.' The concrete foundations are, moreover, 
 continuous in structure, as are the walls themselves, except at one point. 
 The level of the floors also is the same throughout, though the foundations 
 of the court on the north are 75 centimeters higher than the rest.' 
 
 The plan of the new group of rooms is simple. The center of the group 
 is a large hall," out of which open three smaller rooms'" on either side. Be- 
 yond these rooms, and connected with them by large windows, there are 
 two smaller halls," or courts, which, like the central hall, opened directly 
 upon the garden. At the northern end of the group were two small rooms,*^ 
 the entrance to which was from the adjoining court. The new group of 
 rooms, when viewed in its relation to the plan of the Atrium as a whole," is 
 marked by a slight irregularity in position. This irregularity, however, 
 which consists in the location of the central hall almost 3 meters to the 
 south of the main axis of the court on the west, was not so apparent at the 
 
 ' When the level of the east end of the central ' The exact limits of the group on the north and 
 court and of the adjoining rooms on the south have not been recognized by previous 
 
 north was lowered, these foundations were writers. Cf. Richter, /. c, 90. Huelsen- 
 
 chiseled off, except where a later wall had Carter, Roman Forum, 206 and 213. 
 
 been built on top of them. See p. 43, and ' This is especially marked in the almost unbroken 
 plate VI, fig. i. lines of bonding-courses throughout the 
 
 ' Also the foundations of the walls along the sides whole structure. 
 
 of the later court projected originally beyond ' The reason of this change is not clear. It was 
 the walls. due, probably, to local conditions, which 
 
 ' The brickwork of the two periods differs but are no longer apparent, 
 
 slightly. It is, therefore, at times scarcely ' Plan C, 35. 
 
 distinguishable. "Plan C, 32-34, 36-38. 
 
 ' Auer, Der Tempel der Vesta, 9. "Plan C, 31 and 39. 
 
 » Cf. Richter, /. ■.. 90; Huclsen-Carter, Roman "pian c, 29-30. 
 
 Forum, 206 and 213. "Plan C. 
 
THE IMPERIAL ATRIUM OF THE THIRD PERIOD. 31 
 
 time of the erection of the new rooms as later/ since the Atrium was then 
 divided into two distinct parts. The approach to the new rooms was through 
 the garden upon which the central exedra and the smaller courts opened 
 directly. 
 
 The large hall, the center of the new group of rooms, which has been 
 called the tahlinum, though resembling more nearly an exedra, is 8.97 
 meters wide and about 12 meters long.^ In the center of the ceiling of this 
 hall, which is covered at both ends by barrel vaults, was a large opening, 
 for which there was probably substituted in the next period a shaft for the 
 lighting of the upper rooms. The length of the rooms on the north of this 
 hall was 3.55 and of those on the south 3.60 meters. The width' was the 
 same in the corresponding rooms on the two sides. The doors were originally 
 1.62 meters wide and 2.60 meters high. Their height has been, however, 
 somewhat lessened by the raising of the floors, both of the exedra and of the 
 rooms themselves. The windows, which open upon the smaller courts, were 
 all, it is probable, 1.75 meters wide and 2.45 meters high, although several of 
 them have been altered in later restorations. The smaller courts were con- 
 nected with the central part of the new group only by the windows into the 
 small rooms. Of these courts that on the north is 12.06 meters long and 
 8.60 meters wide, while that on the south was smaller, being but 11.80 
 meters long and 6.64 meters wide. The walls of the court on the south 
 have suffered much from later restoration, especially on the south and west, 
 where they have been rebuilt from the level of the later pavement. The two 
 smaller rooms^ beyond the court on the north, the inner one of which occu- 
 pied the site of the ancient altar,^ were but 4. 11 meters long and 5.56 meters 
 wide. 
 
 The purpose of this group of rooms is not clear. Their remoteness from 
 the temple makes any connection with its rites improbable. On account 
 of the number of the smaller rooms' adjoining the exedra, they have been held 
 to be the drawing-room and sleeping-rooms of the Vestals. Although the 
 number, which corresponds to the number of the priestesses, can scarcely 
 be accidental, the isolation of the rooms from the rest of the Atrium with the 
 consequent removal of the Vestals from the protection and assistance of 
 their attendants, renders them unsuitable for such a purpose. They may 
 very well have been, however, the private oflices of the priestesses and the 
 depository for their records and insignia of office. The smaller courts may 
 have been reception-halls or triclinia, such as are often found adjoining the 
 garden in Pompeian houses. 
 
 • See plan E. The destruction of the division wall ' The width is 4.26, 3.99, and 4/58 meters. 
 
 between the two parts of the Atrium first * Plan C, 29-30. 
 
 made the lack of symmetry conspicuous. ' See p. 19. In the wall behind the altar was built 
 'The length on the north side is 12.7 meters, but a small niche, of which mention has been 
 
 on the south side 13 centimeters less. Such made above (p. 19, n. i.) 
 
 irregularities are not uncommon. ' Plan C, 32-34> 3^3^" 
 
32 THE ATRIUM VEST^. 
 
 Architectural Details: The original height of the rooms of this group 
 can not be determined, since the period of the upper part of the walls is 
 uncertain. The height of the exedra is at present 1 1 meters, while that of 
 the rooms adjoining is but a little less. The court on the south, the upper 
 part of which is wholly rebuilt, is 8 meters high. The walls of the court on 
 the north are too much destroyed to allow of the possibility of any decision 
 concerning them. There are no traces in any of the rooms of the original 
 pavements.* They were possibly of opus sectile of the same type as those 
 still preserved in the rooms^ of the same period on the south. For the first 
 time in the history of the building, a system of permanent hypocausts' was 
 constructed. Arched openings* 75 centimeters wide and 70 centimeters high 
 were placed beneath the windows of the small rooms, through which these 
 hypocausts were supplied with fuel. In the rooms on the south of the exedra 
 an upper floor, supported by amphorae cut in halves, was added at a later 
 time as a protection against the dampness. The original decoration has in 
 no case been preserved. The marble posts and wall-facings in the exedra 
 belong to a much later period. No traces of an upper story or of a stairway 
 leading to one are left. The existence of an upper story over the central 
 rooms is, however, suggested by the thickness of the walls. 
 
 The Rooms on the South: The group of rooms on the south' is much 
 smaller than that just described, filling merely the space occupied in the 
 earlier periods by the tablinum." In orientation and level the rooms in 
 general conformed to those immediately adjoining them. The walls are in 
 large part preserved for several meters above the ground. 
 
 Arrangement and Description of the Rooms: The independence of this 
 group of rooms is clear from the lack of any structural connection between 
 them and the rooms on either side and from their dissimilarity in type of 
 construction. In the walls between which the rooms have been inserted, 
 moreover, doors' have been cut into the adjoining rooms for the admission 
 of light and air and to afford greater ease in communication between the 
 various parts of the building. These doors would certainly have been built 
 and not cut, had the rooms under discussion been erected at the same time 
 as those adjoining them or earlier. They must, therefore, have been built 
 at another and a later period. That they form not only an independent 
 
 ^ The pavement of large slabs of fine marble which * The most of these openings have been filled in at 
 still remains in the exedra is much later. a later time; one of them, however, in the 
 
 * Plan C, 13 c and d. The pavement in room 1 1 was north court still remains open, 
 
 inserted, probably, at this time. ^ Plan C, 13 a-d. 
 
 ' The hypocausts of the Atrium do not form a » Cf. plan B, 13 and plan C, 13 a-d. 
 
 single system supplied with heat from a ' The doors have been cut down to the level of this 
 common source, but were arranged in small period, as is shown by the traces of stucco 
 
 groups, which were heated by fires placed on the lower part of the posts. This would 
 
 directly underneath the upper floor. These not have been done, had the doors belonged 
 
 fires were supplied with fuel through openings to the next period, in which the level was 
 
 in the rear of the various rooms. raised. 
 
THE IMPERIAL ATRIUM OF THE THIRD PERIOD. 33 
 
 group but a single structure is evident from their harmony in plan and 
 their structural unity. Their striking agreement, in methods of construction 
 and in the materials used, with the rooms on the east leaves no doubt as to 
 the connection between the two groups. 
 
 The group is modeled after the earlier groups of rooms on the north and 
 west,* not only in its general plan but also in many structural details, such 
 as the width of the doors and the size of the small brick pillars, which become 
 later so conspicuous a feature of the whole building. The group consists of 
 three rooms connected by doors and opening upon a narrow corridor.^ By 
 connecting the newer rooms with the older ones on either side, the whole 
 south side of the Atrium was closely united. The corridor is connected 
 with the larger one outside by a door 1.77 meters wide and by two lofty 
 windows. The door in the rear of the room toward the west may have been 
 cut through at this time, though the reason for it is not apparent. The wall 
 in the rear of the room into which this door opens, in which are the niches 
 for the statues of the household gods, can not have been built later than this 
 period,' since in the next period it was cut off from the room by a second wall. 
 
 Architectural Details: The height of the rooms is not known, but that 
 they were lofty is suggested by the height of the walls still standing. In one 
 of the rooms,* under a hypocaust of the next period, is a fine pavement of 
 opus sectile,^ made of giallo antico, rosso antico, Porta-Santa, pavonazzetto, 
 and other fine marbles. In the corridor are a few pieces of a similar pave- 
 ment. Of the original wall decoration nothing remains, unless it be a dainty 
 bit of fresco' on the side of one of the doors. Though no stairs are left, it is 
 probable that a second story existed over these rooms, as well as over those 
 adjoining. The rooms were probably, from their position, small reception 
 or guest rooms. 
 
 Construction and Materials: The construction of the walls of the groups 
 both on the east and on the south is of the distinctive type which may be 
 recognized ever3rwhere as that of Hadrian.' In the construction of the east 
 rooms the use of a barrel-vault supported by similar smaller vaults on either 
 side has been rightly noted by Auer.' His assumption" that this method of 
 construction is peculiar to the period following the fire of Nero is, however, 
 unwarranted. In both groups the bonding-courses appear regularly and 
 are from twenty-one to twenty-eight courses apart. They are made from 
 
 ' Plan C, 2-7; 21-23. ' "^^ "^^ °^ "P"' reticulatum inclosed between 
 
 ^ The rooms are 4.80 meters long and 5.58, 5.60, bands of brickwork, which is held to be 
 
 3.10 meters wide. The corridor is 3.46 the mark of Hadrian's construction, is not 
 
 meters wide. a certain test. The type of construction 
 
 3 The construction of the wall is peculiar. The is, however, certain. In this point the 
 
 exact period of its erection is at present new rooms are perfectly in agreement with 
 
 difficult to determine. the Pantheon, the Mausoleum, and the 
 
 ' Plan C, 13 c. other more important buildings of the period. 
 
 * See p. 40. * Auer, Der Tempel der Vesta, 6. 
 
 • See Boni, Not. d. Scavi, 1899, 326f. "Auer, Der Tempel der Vesta, 20. 
 
34 THE ATRIUM VEST^. 
 
 tegulcE bipedales of the yellow type so common in floors of the period. The 
 courses of brick measure from 3.30 to 3.70 centimeters and the layers of 
 mortar from 1.50 to 1.75 centimeters. The bricks, which are in part made 
 from roof-tiles, are magenta-red in color and of a finer texture than those of 
 the preceding periods. While the quality is in general good, the bricks have 
 not been properly burned. The mortar is finer than in the last period. The 
 concrete foundations are noticeable from the almost exclusive use of selce 
 as filling. 
 
VI. 
 
 THE IMPERIAL ATRIUM OF THE FOURTH PERIOD. 
 
 The last of the important changes in the Atrium, by which the building, 
 apart from the court, attained its final form, took place under the successors 
 of Hadrian, the Antonines. This change, which represents the fourth stage^ 
 in the development of the building, consisted in the filling in of the spaces^ 
 which had been left on either side of the garden by groups of rooms' opening 
 upon narrow corridors,* above which, as well as above the rooms of the last 
 period, were added a second and a third story.' 
 
 General Description: The earlier part of the building on the west was 
 not changed. At the eastern end the Atrium was increased in extent by the 
 inclusion of the space which had been occupied by the last five shops on the 
 north. On the south the outer wall of the upper stories was pushed back 
 beyond that of the lower rooms, increasing the extent of the building in that 
 direction from 3 to 6 meters. The level of the new groups of rooms" on the 
 north and the south was the same as that of the group on the east. The 
 two corridors, however, through which these groups were entered, as well as 
 the first of the rooms on the north, were 90+ centimeters lower, agreeing in 
 level with the court.' The older rooms on the south were raised by the 
 insertion of hypocausts, though the room at the southwest corner of the court, 
 through which the others were entered, retained its original level.' The 
 general level of the second story above that of the central court is 9 meters. 
 Too little of the third story remains to make any decision concerning its level 
 possible. The walls of the new building are for the most part well preserved, 
 especially on the southeast, where they are still standing as high as the 
 third story.' 
 
 Arrangement and Description of the Rooms: The differentiation of the 
 walls of the new rooms from those of the earlier periods is, except in a few 
 
 ' See p. 7. ' This is evident in the corridor and in the room 
 
 * See plan C. on the north from the level of the hypo- 
 
 'PlanD, 9-12J 32-35. causts. In the corresponding corridor on 
 
 *L. c, 8 and 31. the south, the stairs, which belong to this 
 
 ^ Flan F, a and b. Behind the court on the south period, ascend from the level of the central 
 
 is a series of low rooms (plan D, 24-30) court. 
 
 forming a mezzanino. If these rooms be ' Near the bottom of one of the door-posts, which 
 
 included, the Atrium was at least four was restored at this time, there are remains 
 
 stories high. The stairs leading to the of the painted stucco with which the walls 
 
 fourth story remain, though the rooms of the room were decorated. See plate x, 
 
 themselves have been destroyed. fig. J; for late level. 
 
 " The level of the garden remained unchanged. ''The me^szanino is here included. 
 
 35 
 
36 THE ATRIUM VEST^. 
 
 cases,' made easy by their dissimilarity in methods of construction as well 
 as by their structural independence. This structural independence is 
 especially noticeable at the points of juncture of the new walls with the 
 front wall of the rooms on the east. On the north, where the concrete 
 foundations of the small court project 60 centimeters beyond the face of 
 this wall, the later wall has been built against and on top of them.^ In a 
 similar manner on the south a shelf, which extended along the front of 
 the earlier rooms, has been utilized in building the new walls.' The relation 
 of the walls to each other is most clearly shown by the similarity in archi- 
 tectural plan of the two groups as well as by their structural unity.* 
 
 By the addition of the new groups of rooms on the north and south and 
 by the extension of the front wall of the older court towards the east, the 
 two parts of the Atrium were united into one and the building assumed a 
 more symmetrical form. The earlier independence of the parts was still 
 recognized, however, in the separation between the court and the garden. 
 By the changes just mentioned, as well as by the destruction of the earlier 
 rooms on the north which had been left untouched in the preceding periods, 
 the eastern end of the building was much altered in appearance and the 
 irregularity in the position of the exedra became apparent.' In the older 
 parts of the building on the west, little change took place, except the raising 
 of the level in one or more of the rooms, of which mention has been made." 
 
 The new rooms may, for convenience of treatment, be divided into four 
 groups: (i) the group of rooms on the north,' (2) the group on the south,' 
 (3) the series of half-story rooms in the rear of the court on the south," 
 which we shall call the mezzanino, and (4) the rooms of the second story." 
 
 The Group of Rooms on the North: The group of rooms on the north" 
 was modeled in its general features on that adjoining it towards the west, 
 with which it was united by the narrow corridor which connected the whole 
 series of rooms on that side of the building." Although it has suffered much 
 from later restorations, it is clear that the group consisted of two smaller 
 and two larger rooms opening upon the common corridor. The corridor 
 itself, which is 2.66 meters wide, was connected with the garden outside by 
 a door 2.66 meters in width and by four lofty windows. The two smaller 
 rooms are 4.90 meters long and 3.46 and 4.06 meters wide. The two larger 
 rooms beyond, which are too much injured to be restored with certainty, 
 are now 13.90 meters long and 4.60 and 4.76 meters wide. The two small 
 
 * Owing to the use of similar materials, the original * Plan D. Cf. plan C and p. 30. 
 
 walls of Hadrian are at times difficult to ° P. 27. 
 
 distinguish from the later restorations. 'Plan D, 8-12. 
 
 ' See plate vi, fig. 2. » Plan D, 31-35. 
 
 » At the southeast corner of the later court. ' Plan D, 24-30. 
 
 * The walls form but one structural unit. The "Plan F a. 
 
 two sides are united by the front wall of the "Plan D, 8-12. 
 
 shops in the rear of the buildings (plan D, "The new rooms with the older ones toward the 
 
 m'-m'). yest form practically a single group. 
 
f 
 
 II . 
 
 *■;■ ■ 
 
 4v' 
 
 E. B. V. D. 
 
 Fig. 1. The Penus. 
 
 Fig. 2. The Shrine of the Penates. 
 
THE IMPERIAL ATRIUM OF THE FOURTH PERIOD. 37 
 
 rooms' beyond the north court were also lengthened to 9.50 meters and were 
 connected with the newer rooms adjoining them on the west by a doorway. 
 
 The Group of Rooms on the South: The group of rooms on the south,^ 
 which resembles in its general features that on the north corresponding to it, 
 consisted of three small rooms' opening upon a narrow corridor,* at the 
 end of which was a single large room.' This group was entered from the 
 central court through one of the older rooms, the front part of which served 
 as an entrance corridor both for the new rooms and for the stairway to the 
 upper stories. A small door at the rear of the garden aflForded more direct 
 communication with the rooms at that end of the group.' The three smaller 
 rooms were all originally 5.95 meters long. Two of them, however, were a 
 little later shortened by the insertion of a thin wall in the rear, by which a 
 passage-way was formed, along the front of which a number of openings 
 were left. Through these openings the furnaces of the various rooms were 
 supplied with fuel.' Opening from the corridor upon the garden were two 
 windows high above the ground, through which the rooms as well as the 
 corridor received their light. The larger room, or hall, at the end of the 
 corridor, behind which ran the stairway to the upper stories, also received its 
 light from the garden outside through a lofty window. The older room, 
 which adjoined the newer group towards the west, was much diminished in 
 size by cutting off from it the stairway leading to the second floor. The door 
 which led from this room into the court was reduced in width from 2.50 to 
 1.75 meters, while that which led into the adjoining room on the west was 
 completely closed. By these changes the newer rooms gained additional 
 privacy and all direct communication with the rooms toward the west was 
 cut off. 
 
 At this time the courts at either end of the group of rooms on the east 
 were almost wholly rebuilt. The upper part of the court on the north is 
 wholly destroyed. Around the top of that on the south runs a row of traver- 
 tine corbels similar to those seen in the Pantheon, upon which rested a 
 cornice." On the north side of this court was built a vaulted cellar" 1.96 
 meters wide and 1.70 meters high, which was entered from the inner corridor 
 belonging to the new group of rooms. At the rear of the court was a basin 
 for water'" 3.75 meters long, 1.77 meters wide, and 60 centimeters deep, 
 which emptied into a sewer running to the northwest. At the back and 
 
 ^ Flan D, I3~I4. * In the original plan a smaller door was built 
 
 * Plan D, 31-35. slightly to the east of the present one. This 
 ' Plan D, 32-34. These rooms are 4.56, 340, and door, for some reason, met with such 
 
 4.14 meters wide respectively and 5.95 disfavor that before the completion of the 
 
 meters long. building it was replaced by the other. 
 
 * Plan D, 31. The corridor was 3.24 meters wide ' For further description of these hypocausts, see 
 
 and 13.88 long. pp. 40-41. 
 
 ' Plan D, 35. This room was 8.26 meters long and ' These corbels were not intended to support a 
 6.90 wide. For its original form, see plan roof, as Jordan supposed. 
 
 A, 10. Traces of its earlier front wall are "Plan D, 23 a and plate vii, fig. i. 
 
 still to be found. >°Plan D, 23 b. 
 
38 THE ATRIUM VEST^. 
 
 sides of this basin were five niches for statues/ above which ran the stairs 
 to the mezxanino? In the north court also there are remains of a similar, 
 though smaller, cellar, which was entered from the garden outside by a 
 door cut in the earlier wall. In front of the cellar there was at a later time 
 a small corridor, the level of which was the same as that of the later court, 
 upon which it opened. From this corridor the furnace of the hypocaust 
 underneath the rooms was fed. In the back wall of this court were built three 
 niches for statues. At a later time others were added on the south side 
 between the windows of the rooms. 
 
 The Mezzanino: A half story above the south court were three low rooms,* 
 which were reached by a stairway built above the basin at the rear of the 
 court. Adjoining these rooms towards the west was a low, windowless 
 passage* 2.38 to 2.52 meters wide, upon the walls of which rest those of the 
 upper story. At the eastern end of the passage a room' 2.82 meters long and 
 3. 10 meters high has been at some time cut off from it. Beyond this room the 
 passage-way was at a very late period' partially closed, leaving a furnace-like 
 opening not more than i.io meters wide, inside of which the fire was made 
 for the heating of the caldarium and the rooms connected with it in the story 
 above. 
 
 Purpose of the Rooms: In the rooms on the north but little remains by 
 which their purpose can be determined. From their position and arrange- 
 ment, however, it is probable that they were connected more immediately 
 with the private life of the Vestals, and may very well have been used as 
 triclinia or private reception rooms. It is possible also that the rooms for 
 the entertainment of the guests who sought the protection of the Vestals' 
 may have been in this part of the house. 
 
 In one of the rooms on the south' several structures were discovered in 
 the earlier excavations," which have been generally recognized as the foun- 
 dations of ancient stoves. In the room*" adjoining this on the west there 
 were found also the remains of an ancient mill, and in the court on the other 
 side" the vaulted cellar and the basin for water which have been described 
 above. In the vaulted cellar were found at the same time many broken 
 pottery vessels and three large dolia sunk in the earthen floor. '^ Earlier 
 writers,*^ failing to recognize any difference in the periods of the objects 
 
 ^ See plate vii, fig. z. ^ The Atrium as well as the temple possessed the 
 
 2 Plan D, 24-30. rights of sanctuary. 
 
 ' Plan D, 24-26. These rooms are 5.30 meters « Plan D, 32. 
 
 long and 4.77, 2.97, and 2.66 meters wide. ^ Several of the structures now seen have been 
 
 ' Plan D, 27-30. excavated recently. 
 
 > Plan D, 27. "Plan D, 33. 
 
 " The conversion of the passage-way into a furnace "Plan D, 23. 
 
 took place when the bath-rooms were added '^For a picture of the dolia at the time of their 
 
 in the upper story. The construction is of discovery, see Jordan, I.e., plate xii. 
 
 a late type, as is that of the bath-rooms "Jordan, Der Tempel der Vesta, 64(1. Lanciani, 
 
 themselves. R. and Excav., 232. 
 
THE IMPERIAL ATRIUM OF THE FOURTH PERIOD. 39 
 
 found and misunderstanding their significance, held that the rooms had a 
 religious purpose and were designed for the use of the Vestals in the prepar- 
 ation of the mola salsa, the muries, and the other materials for sacrifice and 
 purification which were committed to their charge. The rooms set aside 
 for that purpose were, however, as has been said, at the other end of the 
 Atrium and were distinct from it. In recent times the rooms have been 
 more commonly held' to be those of a private bakery connected with some 
 one of the many extensive establishments which found quarters in the 
 Atrium after its abandonment by the Vestals. It is very probable that in 
 the last days of the empire the rooms served such a purpose. Since, however, 
 the mill and the greater number^ of the stoves are of a very late period, 
 any such assumption concerning their original use is unwarranted. They 
 were instead, at the time they were built, the center of the domestic life of 
 the household. The room in which the stoves are found was the kitchen, 
 with which the court adjoining it was closely connected. The exact purpose 
 of the room in which the mill was afterwards placed is not clear, but that it 
 also was very closely connected with the kitchen is evident from the presence 
 of a door' between these rooms. The vaulted cellar in the court was, as is 
 plain from its contents, the penus of the household.' The basin near it was 
 for the water needed for the ordinary uses of the household. From their prox- 
 imity to the penus and to the kitchen, one may safely assume that the statues 
 which occupied the niches above the basin were those of the Penates. The 
 court on the north was, in all probability, an open air triclinium. In this 
 court there was, as has been said,^ a second vaulted cellar, resembling in 
 every way that in the other court. Though the existence of more than one 
 penus is peculiar, no other explanation for the presence of this cellar seems 
 possible. It was changed at a later time into a basin for water by the 
 insertion of narrow cross-walls. The niches in the back wall of the court 
 are but three in number. It is probable, therefore, that the divinities wor- 
 shiped here were the Lares with the statue of the ruling emperor between 
 them." The purpose of the large rooms' beyond this court is unknown, 
 though it is possible that the inner room continued to serve as a sacellum 
 for the divinity whose altar had occupied its site in earlier times. The mez- 
 zanino^ adjoining the court on the south' contained the rooms for the slaves 
 of the household. The three larger rooms" belonging to this group, which 
 is entered by a narrow stairway from the court below, may have been 
 
 * Huelsen-Carter, Roman Forum, 212. ®P. 38. 
 
 ' Only two or possibly three of the stoves are early. " In the private cult of the Lares the pater familias 
 
 * This door was blocked up later by the mill and held the central position. The emperor, as 
 
 by a structure on the opposite side. Pontifez Maximus, occupied the same 
 
 * The penus here described must not be confused relation to the Vestals that the pater familias 
 
 with the Penus Festa, which was in the did to the household, 
 
 temple and belonged to the cult. The 'Plan D, 13-14. 
 
 Vestals as a household possessed a penus ''PlanD, 23. 
 
 as well as private Penates. See plate vii, ' Plan D, 24-26. 
 figs. I and 2. 
 
40 THE ATRIUM VEST^. 
 
 originally intended for storerooms, or have formed a dark passage* like that 
 adjoining them towards the west. 
 
 The Shops: At this period the street on the north was spanned by arches 
 resting on concrete pillars, which were built against the travertine posts of 
 the first period.^ By the addition of these pillars the street became 1.80 
 meters narrower. The level both of the street and of the shops was raised, 
 especially towards the east, where the street is 2 meters above the level of 
 the republican remains below.' On the east a new wall was built in front 
 of the earlier shops.* 
 
 Architectural Details: The height of the rooms on the north can not be 
 determined. The corridor and rooms on the south, however, which were 
 covered by vaulted ceilings, were 7 + meters high. The archways by which 
 the rooms were connected with the corridor were but a little lower.' Though 
 the upper floors of the hypocausts, upon which the pavements were laid, 
 are still left in many of the rooms, the pavements themselves have almost 
 wholly disappeared. Those found in several of the rooms on the south are 
 of a later period. A part of the pavements which are preserved in the mez- 
 zanino may be, however, original. In the older room' on the south in which 
 was built the stairway leading to the upper stories, there is still preserved a 
 fine opus sectile pavement, which is often assigned to this period, though it 
 belonged more probably to that of Hadrian.'' 
 
 In the room on the north adjoining the entrance-corridor is a hypocaust 
 55 centimeters high, which was built at the same time as the rooms them- 
 selves.' Hypocausts of a similar type were, it is probable, built under the 
 other rooms, though at a higher level.* In the new rooms on the south, also, 
 hypocausts were built throughout, raising the level of the whole group to 
 that of the group on the east. In several of the rooms the double floors of 
 the hypocausts are still preserved. In the corridor, unfortunately, they have 
 been destroyed, though traces of them still remain. In the room farthest 
 toward the east a second hypocaust has been placed on top of the earlier 
 
 ' The door and windows do not belong to the ' From the traces which remain it is probable that 
 
 earliest period, but were cut out of the the hypocaust under which the pavement 
 
 wall later. was found is not later than this period. 
 
 2 Cf. plan C and see pp. 18-19. The pavement itself, therefore, must be 
 
 3 See plate viii, fig. i . The level of the street is earlier. In type also it agrees perfectly 
 
 shown by the height of the concrete founda- with the pavements which are usually as- 
 
 tions of the pillars, which were added at this signed to the time of Hadrian, 
 
 time. <In building the walls of this room, the bonding- 
 
 * Plan D, m''-m". course which is at the level of the lower floor 
 
 ° The archway into the kitchen (plan D, 32) was, of the hypocaust has been so laid as to allow 
 
 possibly before the completion of the build- the ttgula bipedales to project beyond the 
 
 ing, replaced by a small door. Traces of wall. On the shelf, so formed, has been 
 
 the earlier archway are, however, still to be placed the outer row of supports for the 
 
 found, though its width can not be deter- upper floor of the hypocaust. 
 
 mined. "The hypocaust still to be seen in one of these 
 
 « Plan D, 36. rooms is of a much later period. 
 
Fig. 1. The Street on the North of the Atrium. 
 
 E. B. V. D. 
 
 Fig. 2. Remains of the Upper Stories. 
 
THE IMPERIAL ATRIUM OF THE FOURTH PERIOD. 41 
 
 one, by which the level of the room was raised above that of the corridor. 
 In the wall in the rear of the two smaller rooms* arched openings were 
 left below the level of the pavement, through which the furnaces placed 
 beneath the upper floor of the hypocausts were fed. In the rear of the 
 larger room toward the west,^ and adjoining one of these openings, are the 
 partly destroyed remains of one of these rude furnaces. Traces of others 
 are found in various parts of the Atrium. Hypocausts were at this time 
 inserted also into the older rooms on the south, with the exception of two' 
 which retained the level of the outer corridor on which they opened directly. 
 In connection with these changes the doors between the smaller rooms^ 
 belonging to the last period were closed, and those by which they were 
 connected with the corridor were reduced in width. The older rooms' on 
 either side of this group, the alee of the earlier house, were at least partially 
 cut off from the court, upon which they opened through wide arches. In 
 one of them* a wall 45 centimeters thick was built, in which is seen a hypo- 
 caust opening similar to those in the newer rooms toward the east. The 
 shrines of the household gods,' which by the insertion of the wall had been 
 made inaccessible except through a narrow passage-way, were abandoned 
 and the statues of the gods were removed, probably to the court* at the north- 
 east corner of the Atrium, in the back wall of which niches were at this time 
 built. In the passage-way just mentioned was discovered in 1899 a hoard 
 of coins of the late empire." In the large room on the west of the court" 
 hypocausts were inserted or the earlier ones entirely rebuilt, since the tegulcs 
 bipedales, of which the floors of the hypocausts are made, are almost wholly 
 of this period. In several of the rooms there are scanty remains of frescos, 
 which may possibly belong to this period. It is more probable, however, 
 that they belong to the final restoration of the building in the time of Sep- 
 timius Severus." Marble also was used extensively in many parts of the 
 building. 
 
 The Upper Stories: Of the upper stories of the Atrium, as a whole, but 
 a few rooms remain, which are situated at the southeast corner of the build- 
 ing.*^ These rooms belong entirely to the period under discussion. The 
 further extent of the rooms of the period can not be determined, though it 
 is certain that there was a group on the north corresponding to that on the 
 south. Of the stairways" leading to the group on the north only the sup- 
 porting walls are left. The stairway" on the south divides at a short distance 
 
 ' Plan D, 33-34. ° ^"'^ further description of these coins, see Not, 
 
 2 Plan D, 35. d- Scavi, 1899, 3278. 
 
 ' Plan D, 36 (front part of room) and 43. '"Plan D, 51. 
 
 * Plan D 38-40. "Too little remains to make any decision concern- 
 
 "PlanD, 37 and 41. ing them possible. 
 
 6 pijn j)^ ^2. "'Plan F a; see plate viii, iig. 2 and plate ix, fig. 1 . 
 
 7 See p. 33. "Plan D, adjoining 9 and 13. 
 
 ' The number of niches in both rooms is the same. "Plan D, 35-37- 
 
42 THE ATRIUM VEST^. 
 
 from the bottom. The part towards the west has been almost wholly 
 destroyed/ although that leading to the rooms on the southeast^ is still 
 preserved. 
 
 The rooms of the second story correspond in general to those of the lower 
 floor and of the mezzanino. Since, however, the walls are somewhat narrower 
 than those below, to which they correspond in position, the rooms are a 
 trifle larger. On the north side of the narrow corridor in which the stairway 
 ends are three large rooms' looking out over the garden. The rooms on the 
 south* are much smaller and have been raised to the level of those opposite 
 by the insertion of double floors.^ At the end of the corridor are three other 
 rooms" which, like the mezzanino, are lighted from the open court. These 
 rooms were connected directly with the Nova Via by a small door. At the 
 rear of the court a passage-way was thrown across to the opposite side, where 
 there are scanty remains of a stairway' leading to the third story and of one 
 room,' both of which seem to be of a later period. It is possible that this 
 part of the building was originally occupied by a terrace. Behind the exedra 
 are the remains of a passage-way leading to the rooms on the north. 
 
 The height of the rooms is about 5 meters. The pavements which remain 
 belong probably to a later period, since there are traces of others at a lower 
 level. Beneath all the rooms, except the three next to the garden, are hypo- 
 causts of a later period. It is probable that the double floors of the original 
 building were the same height, since the level of the rooms would otherwise 
 have been below the level of the corridor. The group of rooms was probably 
 that in which were the private apartments of the Vestals. The smaller rooms 
 may have been used for baths as was the case later. 
 
 Construction and Materials: The walls of the new rooms agree in con- 
 struction with those elsewhere recognized as of the period of the Antonines. 
 The walls of the lower story are 89 or 75 centimeters thick, while those of 
 the upper stories are either 75 or 60 centimeters. There are no bonding- 
 courses in any of the walls.' The materials used do not differ from those 
 of the preceding period. 
 
 ' The supporting wall has, however, been pre- ' It is possible that these were intended, when 
 served and is identical in width and con- built, for hypocausts. There are no traces 
 
 struction with that corresponding to it on of any means for heating them, however, 
 
 the other side. A second wall exists behind until later, 
 
 this supporting wall, the purpose of which "Plan F a, 9-11. 
 
 is not clear. ' Plan F h, 13 h. 
 
 ' Plan F a, i. s plan p j^ 13 a. 
 
 ' Plan F a, 3-5. » This is the most striking difference in construction 
 
 * Plan F fl, 6-8. between this and the preceding period. 
 
VII. 
 
 THE IMPERIAL ATRIUM OF THE FIFTH PERIOD. 
 
 In the time of Commodus the Atrium was once more partly destroyed by 
 fire. The restoration following this partial destruction marks the last stage 
 in the development of the building.^ In this restoration the rooms on the 
 west and northwest,^ which had suflFered most severely, were rebuilt, and the 
 central court, which had remained unchanged during the previous periods, 
 assumed its present form. 
 
 General Description: Since the fire and the reconstruction following it 
 affected only the upper portion of the walls, there was no change in the extent 
 of the Atrium. Several important changes took place, however, in its level. 
 In consequence of the extension of the colonnade, the level of the garden was 
 lowered to that of the court on the west, although the level of the colonnade 
 itself was at this time, or a little later, raised 20 centimeters.' Various other 
 changes occurred in the level of individual rooms. 
 
 Plan of the Interior and Description of Changes: The determination of 
 the walls and portions of walls which belong to this period is rendered easy 
 by the distinctive type of their construction.^ The walls which are wholly of 
 the period are, moreover, clearly distinguished from those of the earlier 
 buildings by their structural independence. 
 
 The general plan of the building remained unchanged. Its appearance 
 was, however, much altered by the destruction of the walls' separating the 
 earlier court from the garden and by the extension of the colonnade to the 
 rooms on the east. By the extension of the colonnade the building was more 
 closely bound together and the differences in the height and the style of 
 architecture of its various parts concealed. The colonnade, which was 
 probably wholly rebuilt, consisted, like that of the earlier period, of two 
 rows of columns, one above the other, but with no floor corresponding to 
 the intermediate entablature. The intercolumnar spaces are wider than 
 those of the older part of the colonnade towards the west.' The increase in 
 width, however, is not regular^ and arose from the necessity of distributing 
 the six columns on either side of the newer part of the colonnade over a space 
 
 ' See p. 8. " Schulze (Jordan, /. c, 37) held this increase to 
 
 ^ Plan £, 48-56^, I £2-7. be an architectural refinement due to regard 
 
 ' The new level was that of the coarse black and for perspective. C/. p. 23. 
 
 white mosaic, of which considerable remains ' The intercolumnar spaces on the west do not 
 
 are found. exceed 343 meters. Those toward the east 
 
 ^ For this type of construction, see plate 11, fig. z. are 3.93, 3.63, 3.63, 3.63, 3.73 meters re- 
 
 ' See plate iv, fig. i . spectively, 
 
 43 
 
44 THE ATRIUM VEST^. 
 
 determined by the limits of the earlier building. The octagonal structure,^ 
 the remains of which may still be seen in the center of the court, was prob- 
 ably built to conceal this irregularity. The columns of the lower story of 
 the new colonnade were of cipollino, 50 centimeters in diameter and not over 
 3 meters in height. Those of the upper story were of breccia corallina, 40 
 centimeters in diameter and 2.93 meters high.^ In connection with the ex- 
 tension of the colonnade, the large basin,' which was no longer in the middle 
 of the court, was destroyed and its place taken by two smaller ones* at the 
 ends of the building. At this time or a little later* was built, over the older 
 basin, the octagonal structure mentioned above. The exact character of this 
 structure is uncertain but it was probably an open summer-house, or pavilion, 
 designed to conceal the irregularity in the colonnade and to afford more 
 privacy to the rooms set apart for the daily life of the household. 
 
 No change took place in the size or arrangement of the rooms of the 
 Atrium proper. At a somewhat earlier period the large room* on the north 
 next to the temple precinct had been divided into three smaller rooms open- 
 ing directly upon the area in front of the temple. At this time a similar 
 change was made in the room' on the south of the area. In the room* adjoin- 
 ing that in which stood the basin for the washing of the sacred utensils was 
 built a platform 2.15 to 2.84 meters wide and 1.08 meters high, the purpose 
 of which is not clear. It may have been connected, however, with a tem- 
 porary stairway thrown across to rooms which no longer exist. Beside it 
 a narrow stairway was built opening directly upon the ramp leading to the 
 Palatine. By this stairway an easier means of communication was provided 
 with the fountain of Juturna, from which water was probably brought for 
 the uses of the cult. Underneath the platform are the so-called ovens, which 
 resemble in form the modern Dutch ovens. They were intended for the 
 storing away of articles used in the sacred rites or for the ashes of the focus, 
 which probably stood near." That they were not ovens is clear from their 
 form and construction'" as well as from the absence of any opening for the 
 escape of the smoke or the creation of a draught. There are, moreover, no 
 traces of the presence of fire inside of them. 
 
 The Nova Via was at this time spanned by arches 4-I- meters high sup- 
 ported by pillars 1.47 meters square, which formed an arcade extending 
 
 ■ Flan £, u. The top of this basin has been ' Bricks bearing stamps of a later period have been 
 
 restored. found in the structure. They belong, how- 
 
 ' Several columns of breccia corallina are still pre- ever, to a restoration, since the construction 
 
 served. Only fragments remain of the ci' in general is of the time of Severus. 
 
 poUino columns, which resisted less well the ' Plan D, i. Cf, plan E, la-ic. 
 
 action of fire. Fragments remain also of ' Plan D, 56. Cf. plan E, s6a-$6c. 
 
 the entablature of white marble, as well as ^ Plan E, 54. 
 
 a number of Corinthian capitals of various " This was first suggested to me by Professor Mau. 
 
 sizes. '"See plate IX, fig. 2. For the form and construction 
 
 " Plan D, O. of ancient ovens, see Mau-Kelsey, Pompeii, 
 
 •Plan E, r and s. These smaller basins, although 391. Such places for the stowing away of 
 
 they are of the same period, are not of the various objects are often seen in Pompeian 
 
 same size. houses. 
 
E. B. V. D 
 
 Fig. 1. The Street east of the Atrium. 
 
 Fig. 2. The So-called Ovens. 
 
THE IMPERIAL ATRIUM OF THE FIFTH PERIOD. 45 
 
 from the corner of the Atrium to the juncture of the Nova Via with the Clivus 
 Sacer} A similar but smaller arcade was built along the narrow street in the 
 rear of the Atrium. 
 
 Architectural Details and Construction: The height of the rooms can in 
 no case be determined. In the court are pavements of several periods. Of 
 these that of opus spicatum belongs probably to this period,^ though that of 
 the coarser black and white mosaic is but a little later. Many of the pave- 
 ments oiopus spicatum throughout the house are, however, of a later period. 
 The pavements oi opus sectile made of large slabs of marble, which are found 
 in the court and in many of the rooms, are also of a later time. The remains 
 of decoration which are found in many parts of the Atrium belong in great 
 part to this restoration. The most striking of these remains is the dcido of 
 marble, with which the court and a number of the rooms' are decorated. 
 Above this dado the walls were covered with marble facings or with frescos. 
 Traces of both these styles of decoration have been found in various rooms. 
 
 The walls are, with a few exceptions, superimposed on those of the earlier 
 periods, with which they agree generally in width.^ The height to which 
 they have been left standing varies from a few centimeters to several meters. 
 The type of construction differs greatly from that of the earlier walls.' The 
 tegultB bipedales appear, though less regularly than in the walls of Hadrian. 
 The courses of brick measure from 2.75 to 3.25 centimeters in width and the 
 layers of mortar from 2.25 to 2.50 centimeters. The bricks themselves are 
 well burned and of a good quality. They vary in color from yellow to yellow- 
 red and are of a coarse texture, resembling most nearly those of the periods 
 of Nero and the Flavians. 
 
 Later History of the Atrium: Of the changes which took place in the 
 Atrium after the time of Septimius Severus, two only are of importance in 
 its developmeht. The first of these, which affected the Sacellum Larum 
 rather than the Atrium itself, consisted in the construction of two stairways, 
 one a short time after the other, in one of the smaller rooms" connected with 
 the Sacellum and in the erection of a thin wall in the rear of the court, by 
 which the whole group was separated from the rooms of the Atrium adjoin- 
 ing it. In this wall were placed the three niches for statues mentioned 
 above. ^ At some later time another stairway connecting the Atrium with 
 the Nova Via was added in the room' at the southwest corner of the cen- 
 
 * See plate x, fig. i . * The walls of the small rooms on the west which 
 ^ The pavement was laid on the top of one of the are built from the foundations are but 45 
 
 walls across the middle of the court (plate centimeters thick. 
 
 IV, fig. 1), which were broken down in this ^ For the construction of this period see plate 11, 
 
 period. fig. i. 
 
 ' Door posts as well as a dado of marble are found ' Flan £, 44. 
 
 in the exedra. They may have been com- ' P. 25. 
 
 monly used throughout the house. ' Flan £, 43. 
 
46 THE ATRIUM VEST^. 
 
 tral court, the level of which was also raised.^ Underneath the stairway 
 a small shrine was built. 
 
 The other and more noticeable change occurred in the second story, 
 where a system of bath-rooms was inserted in the rooms of the earlier build- 
 ing.^ The rudeness of the construction and the awkwardness of the arrange- 
 ment point to a period later than that of the Vestals. In the rooms next to 
 the stairs on the south is a caldarium^ with a tepidarium* adjoining it. The 
 method of entrance into the caldarium is not clear, since the basin occupied 
 the space next to the door. The frigidarium, if one existed, was in the room' 
 on the other side of the caldarium. In the walls of all these rooms flue tiles 
 are found, by which the heat was conducted from the furnace in the dark 
 passage below." In the first of the larger rooms beyond' a smaller caldarium 
 has been built, which received its heat from a furnace underneath the basin. 
 The water, however, must have been heated in some other manner, since 
 sufficient heat for this purpose could not have been received from the furnace 
 underneath, on account of the thickness of the concrete floor of the basin. 
 The rooms toward the east,' which also were heated both by hypocausts 
 and by flue tiles in the walls, were probably used for a tepidarium and for 
 dressing-rooms. Behind this group of rooms' a corridor was built, from 
 which the furnaces underneath the rooms were fed. Gn the east of the court 
 is a frigidarium^" 2.95 meters square and 1.15 meters deep, which is lined 
 with rare marbles. It was entered by steps leading down directly from the 
 passage-way which connected the rooms on either side of the court. Near 
 the entrance, 1.90 meters above the ground, were holes through which the 
 water poured into the basin below in jets, as in certain houses in Pompeii. 
 At the southeast corner of the building, in one of the earlier shops" opening 
 upon the street in the rear of the Atrium, a tank" was placed, from which the 
 various baths were supplied with water. In the adjoining room a smaller 
 tank was built, near which may have stood another basin. Throughout the 
 whole Atrium many unimportant changes occurred, as the insertion of hypo- 
 causts in several of the rooms" on the north, and the restoration of the older 
 pavements. The colonnade also was at some later time superseded by a 
 brick wall pierced with windows and the corridor on the south was cut into 
 small rooms. 
 
 Of the history of the Atrium in the centuries following 382 A. D., when 
 the building was abandoned by the Vestals in consequence of the decree of 
 
 'See plate x, fig. %. The pavement to which '"Plan F, 156. The frigidarium has been built 
 
 the modern steps ascend is of this late above one of the shops opening upon the 
 
 restoration. street behind the Atrium, into which there 
 
 ^ Plan F b. had been inserted a stairway leading to the 
 
 ' Plan F b, 7. upper stories of the house. 
 
 * Plan F b, 6. "Plan E, m'. The shop furthest to the south. 
 
 ■ Plan F b, 8. '*The remains of this tank are very scanty. It 
 
 " Plan E, 28-29. rested on a brick-faced concrete vault 
 
 ' Plan F b, 9. similar to that which supported the adjoining 
 
 » Plan F i, 10-11. frigidarium. 
 
 "Plan F b, 9-11. "For example, see plan E, 11, iz. 
 
THE IMPERIAL ATRIUM OF THE FIFTH PERIOD. 47 
 
 Gratian, our knowledge is very slight. It is probable, however, that it served 
 for several centuries as a residence for the officials of the imperial court, to 
 one of whom doubtless belonged the treasure of the three hundred and ninety- 
 seven coins found in 1899 beneath the pavement of the corridor on the south.' 
 The greater number of these coins,^ which belong wholly to the fourth and 
 fifth centuries, are of the emperor Anthemius (467-472 A. D.). With the 
 passing of time, as the building fell into decay, the poorer classes found 
 shelter within its walls in huts built among the heaps of rapidly accumulating 
 debris, while the marbles of the walls and the statues furnished a source of 
 supply for at least two limekilns located in the building itself.' At this time 
 probably the pedestal dedicated to Coelia Claudiana, which was found in 
 1868 in the stadium on the Palatine, was removed from the Atrium.* That 
 the Atrium was not yet wholly abandoned in the tenth century is shown by 
 the remains of a small house in the northwest corner of the court, beneath 
 the pavement of which in 1883 a vase was found containing over eight hun- 
 dred Anglo-Saxon coins° of the ninth and tenth centuries. From the fibula 
 found with these coins, which bore the name of the pope Marinus, it is clear 
 that the owner of this house was an official of the papal court. After this 
 time the Atrium Vestae was, so far as can now be ascertained, not only wholly 
 abandoned, but its site was forgotten until 1883, when in the course of 
 the excavations it was once more brought to light. 
 
 * Plan E, 42. 3 Lanciani, Not. d. Scavi, 1883, 485. 
 
 ' For a description of these coins, see Not. d. Scavi, * This base has within a few years been restored 
 
 1899, 3278. The coins are now in the to its original place in the Atrium. 
 
 Museo delle Terme. ^ These coins are now in the Museo delle Terme. 
 
PLAN A 
 
 NOVA 
 
 V I A 
 
 m' ;'; m K m ;5 m' K f"' i| m' '- in' A m' fj m' 
 
 Walls, existing or traceable, of 
 the period of Nero 
 
 Walls not existing above ground 
 but practically certain 
 
 Walls seen when excavated but 
 not measured 
 
 .AJfocn &Ca Lilh. Jta/fimorr-^ ^fd. 
 
 cell tinietei'^l meter 
 
 ''•"■^■■'•1 Probable walls 
 
 I I Possible walls 
 
 1 I Walls of the republican period 
 
NOVA VIA 
 
 PLAN B 
 
 m m 
 
 
 
 ■My- 
 
 |o a 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 2 
 
 i ^ i 
 
 
 m 
 
 M '" 
 
 1 m 1 
 
 
 Walls of the Flavian period 
 Walls of the period of Nero 
 CSZZ3 Possible walls 
 
 Republican walls 
 
 .iftiffiiKt'nlitf, noltwttitr.^fU 
 
 \ I'ciii luu'tei'-l nu'ter 
 
I i I 1 J 
 
 NOVA VIA 
 
 PLAN C 
 
 : jj;j 
 
 Walls of the period of Hadrian 
 
 Walls earliei than the period of Hadrian 
 
 Walls of the period of Nero 
 ■■/yy:"- ■■ Walls of the Flavian period 
 
 Probable walls 
 S:SS>;ss; Republican walls 
 
 .III,. ~J ( .. /,//, llullmui,-, W 
 
t_ — 
 
 I ' • 
 
 I 
 
 ii u ^ ^ u ii ^ 
 
 [Hr 
 
 VVV-T^ 
 
 PLAN D 
 
 NOVA VIA 
 
 Walls of the Antonine period 
 Walls of the period of Nero 
 13 Walls of the Flavian period 
 H Walls of the period of Hadrian 
 
 ^SD Probable walls 
 ^'^.y?,;?! Republican walls 
 
 AHf>fnf: r >, ///■/, fJi,ltut,on'.. Hr/. 
 
 V4 i'enrlniet.ei'=l iiit^ier 
 
 H 
 
PLAN E 
 
 i-"^l^li-i!^i"W 
 
 a a 
 
 a iL 
 
 NOVA 
 
 V I A 
 
 ,;,//; III, limn. I- ^1'' 
 
 % i'p|itlnif'ti'i'--l m^rer 
 
 Walls of the peiiod Septimius Severus 
 Walls of uncertain date earlier than the 
 
 period of Septimius Severus 
 Walls of tlie period of Nero 
 t :V^,.v< Walls of the Flavian period 
 
 Walls of the period of Hadrian 
 
 Ki'';:'. I Walls of the Antonine period 
 
 Walls of the period of Septimius Sevens 
 ^^^^ resting on earlier walls 
 ^^^ Walls later th;in the period of 
 Septimius Severus 
 
 Republican walls 
 
PLAN F 
 
 ^■■H Late walls 
 
 I I Walls of the period of Hadrian 
 
 Walls of the Antonine period