Number 90

Numismatic Notes and Monographs is devoted to essays and treatises on subjects relating to coins, paper money, medals and decorations, and is uniform with Hispanic Notes and Monographs published by the Hispanic Society of America, and with Indian Notes and Monographs issued by the Museum of the American Indian—Heye Foundation.

Publication Committee
• Agnes Baldwin Brett , Chairman
• Stephen H. P. Pell
• Thomas O. Mabbott

Editorial Staff
• Sydney Philip Noe , Editor
• Sawyer McA. Mosser, Associate Editor

copyright, 1940, by

the american numismatic society

the merrymount press, boston, u. s. a.

∵

THE serious use of coins as a means to facilitate the study of ancient architecture started with T. J. Donaldson's Architectura Numismatica, published in 1859. A casual selection of types was discussed, but the book today has little value beyond that of curiosity. A similar selection of types was used by H. Jordan in his monumental Topographie der Stadt Rom, 1885, and in his monograph on the temple of Vesta which appeared in 1886. Furtwängler and others at the same time were using coins as very minor aids to scholarship. No really scholarly numismatic-architectural studies appeared until the publications of H. Dressel at the turn of the century, which included researches upon the ludi saeculares series of Domitian, 1 the temple of Vesta, 2 the temple of Venus Cloacina, 3 the temple of Diva Matidia, 4 and the temple of Isis Campensis. 5 Herein the first attempt was made to collect die-varieties and to evaluate the information to be found upon coins as far as the actual buildings were concerned. Once the direction in which research might go had been pointed out, numerous other scholars followed Dressel's lead. W. Weber studied the coin types showing the temples of Mercury and of Isis Campensis. 6 Katherine Esdaille published a small article on the temple of Magna Mater in Rome. 7 O. L. Richmond wrote about the temples of Concord and Apollo Palatine. 8 More recently, in works too numerous to list here, the value of the architectural representations upon coins has been increasingly appreciated. General studies of architectura numismatica have more or less cleared the field for more detailed and specialized studies. 9 In none of these publications, however, is there a very detailed discussion of methods of approach to the whole study of architectural representations.

The main body of the following discussion consists of the broader results only of a research problem which has occupied some five years of study. During this time all available publications of Roman coins have been combed for references and reproductions. The major collections of Europe and America have been canvassed for material—an undertaking which the present conflict has unfortunately greatly hindered. Altogether over 2000 pertinent coins have been examined and listed. Since a certain amount of significant variation was found to exist between different dies of the same issues, it was decided to collect all the available die-varieties for each issue. This work is still under way, and it is hoped that the resultant Corpus will some day be published along with the reasoned arguments which have led to the results which follow.

The types chosen for this study have been only those for which some reason could be advanced to assume that the structure represented was in the city of Rome. 10 The minting place of the pieces helped considerably in arriving at such a conclusion, since it was considered reasonable to claim that the mint in the city of Rome would seldom strike coins showing extra-city temples, and conversely that mints outside of Rome would have little reason to commemorate temples in the city of Rome upon their coins. Contradictions of the latter statement are found, but in very limited number, and only from mints which were parts of the imperial system. 11

As an example of coin types showing provincial temples, although minted in Rome, only one need be cited to establish the category, i.e., the temple of Venus Erycina in Sicily (B.M.C.R. Pl. XLVII, 21). This type and others like it will be discussed in projected studies of the various geographical divisions of the Empire.

When there is no inscription on the coin to aid in the identification of the building shown, it is necessary to refer to the ancient literary sources and to endeavor to equate the buildings thus shown with those mentioned in the sources for the emperor or moneyer in question. This method is not a touchstone for solving the problems of all the uninscribed temples, but a few were identified which had not been before. 12 Those which still defy identification with a site or at least a specific divinity appear in the list of Appendix A as "distyle 'A' and 'B,'" "tetrastyle 'A' and 'B,'" "hexastyle 'A' and 'B,'" and "VOTA."

The chronological limits within which Roman temple types appear upon Roman coins are 88 B.C. ( Jupiter Capitolinus of the gens Volteia) and 311 A.D. (Venus and Rome of Alexander Tyrans). During this period, forty-three different temples of Rome are used upon the coins. A study of these as they appear in Appendix A will reveal that architectural representations came into favor in Roman coinage only gradually. In the entire first century B.C., before the reign of Augustus, only eight moneyers use them. With Augustus and in the first century A.D., the tempo quickens. Augustus himself strikes three varieties, Tiberius two, Caligula one, Nero two, Vespasian four, Domitian seven, Titus three. In the second century A.D., Trajan issues five varieties, Hadrian four, Antoninus Pius seven, Marcus Aurelius five. With the third century the incidence of varieties drops off, and most emperors who strike coins with architectural temple types are responsible for only one or two varieties. Many do not strike them at all. The fourth century sees only one variety issued, the Venus and Rome temple (cf. Pl. IX, 1, 3, 5–8). 13 Indeed, from the reign of Septimius Severus onward, the temple of Venus and Rome is the only one consistently figured upon coins and medallions. A number of temples are occasionally figured, but limited generally to a single issue of a single emperor. 14 With four exceptions, Genius Exer- citus, Jupiter Sospitator, Vesta in Foro and Jupiter Ultor, these occasional types are found only upon medallions and not on coins. The first two and the last of these exceptions are very restricted in use, being confined to issues of single emperors. Vesta in Foro, on the other hand, receives a use which approximates that given to Venus and Rome , although the disparity in the frequency of occurrence (Vesta in Foro occurs on seventeen late issues and Venus and Rome on over one hundred) makes any comparison between the usages of these two types of doubtful value. However this may be, these two types are the dominant ones of the late third and early fourth centuries A.D. upon coins of Roman mintage. The type of Juno Martialis (Pl. VIII, 1), which achieves a certain popularity under Trebonianus Gallus and Volusian, and appears neither before nor after their reigns, must be regarded as an exception in the otherwise clear picture of numismatic temple iconography of this late period. For some as yet unknown reason, these two emperors felt particularly beholden to this goddess, and struck a considerable number of issues in her honor.

All sorts of structures from a lararium (Minerva, Pl. IV, 1) to a complex of temples and porticoes (Diva Matidia, Pl. V, 3) have been used as models for the various types. Although it is obvious that there must have been some compelling reason for the selection of any one type, it has not been possible to arrive at a blanket rationalization for the various choices. However, the use of most of the types can be explained by some one of the reasons which follow.

The most usual reason appears to have been actual building activity of the moneyer or emperor. In many instances, our sources are too scanty to assign definitely the use of a type to this reason, but it seems to be assured in the majority of instances. 15 Some of these types, as will be developed later, were chosen for other reasons also. However, definite issues for all of the above can be found, which may be equated with building activity on the particular temple.

Closely connected with this group, chosen because of active building, are the Neptune (Pl. I, 5) coins of the gens Domitia and the Dims Julius (Pl. I, 7) coins of Augustus. Considerable discussion has been aroused by both of these issues because their date of emission in each case is earlier than the known dedication date of the temple shown upon the coin. An iconographical peculiarity, hitherto unnoticed, also separates these types from the general run of architectural representations. Both these structures, it is to be noted, and only these in the whole series of temple types, are shown placed upon a rather high podium which has no visible means of access to the floor level. This common peculiarity, coupled with the common trouble found in dating these two issues, makes it obvious that some common fact is being expressed by the iconography of the representations. This fact is that neither structure was built at the time the coins were struck, and the inaccessible podia are meant to indicate that only putative elevations are shown on the coins and not real structures. Therefore, as showing structures projected, even though not yet built, these two coins belong in the group of types chosen because of building activity.

A second reason for the choice of a temple type is the commemoration of some event of extraordinary im- portance to the Empire. This can be used to explain the use of four types. 16 There may be many other types chosen for this purpose, but our knowledge does not allow us to isolate them. In none of the cases noted does actual building activity appear to have been involved, and for this reason the group seems a valid one, especially since there is an event which explains each issue.

A third reason is the pictorial commemoration of the celebration of religious ceremonies. 17 The group has been isolated through the common factor of a religious action going on before the temple.

A fourth reason is the pictorial commemoration of an essentially political event. The Divus Julius (Pl. V, 1) series of Hadrian, showing the Emperor addressing citizens from the rostra before the temple, is the single instance which comes under this heading.

It will be observed that a number of types do not fall into any of the above categories, such as, for instance, Genius Senatus of Antoninus Pius, Juno Martialis of Trebonianus Gallus, etc. Probably if our knowledge of things Roman were more extensive, these too would be found to fit into one of the categories just described.

With a few exceptions, it is possible to check the details of the coin types with literary sources concerning the various buildings and with the actual remains themselves. The agreement of the coin types in all major details with the sources is striking, and the necessary con- clusion is that only actual buildings are shown as coin types (with the exception, of course, of the two putative elevations seen on Divus Julius of Augustus and Neptune of the gens Domitia), and these with considerable accuracy of observation. Certain necessary abbreviations are used, but the essential features of the buildings are faithfully reproduced.

In those cases, for instance, where we know that the building shown was round, it is without exception represented as round upon the coins, e.g. Bacchus (Pl. VI, 1), the two temples of Vesta (Pl. II, 3, 4; III, 3, 4; V, 5), and Mars Ultor (Pl. I, 6, 11). There appears not to have been a set iconography for round temples, but the number of columns of the peristyle never exceeds six on the numismatic representations, even in those cases where we know that twenty was the actual number in the building. A minimum of two is also found, in which instances the roundness of the building, indicated by the position of the columns on those coins where more than this number appeared, is now shown by the curvature of the steps or by the arrangement of the roof. In other instances, where we are informed concerning the façades and number of columns in the porch of rectangular structures, the coin types generally preserve the actual number of columns and also the shape of the façade in its major details. The temple of Concord (Pl. II, 1) of Tiberius is a splendid example of this adherence to actuality. The façade of this temple was of a peculiar extended form, which is easily seen to be the case on the coins. The earlier series of Hadrian and Antoninus Pius which show the temple of Venus and Rome (Pl. IV, 5; V, 2), the Trajanic Venus Genetrix series (Pl. IV, 4), and the Antonine Diva Faustina group (Pl. V, 4) are all examples of this same adherence to actuality. This is, however, not a hard and fast rule, and upon both early and late coins the number of columns of a particular building shown may be at variance with what is known about the structure archaeologically. It is, moreover, important to note that, as far as can be ascertained, columns were never added in the coin types above the real number, so that if a type is found showing, for example, a tetrastyle structure, it is certain that the building as it stood had at least four columns on its façade and perhaps more, but certainly no less than four. A good example of this reduction of columns for reasons of space or composition is offered by the late third century A.D. series of coins showing the temple of Venus and Rome (a decastyle structure) where the number is reduced on the coins to six usually, but occasionally to eight, four, or even two. As far as the architectural orders of the columns are concerned and the presence or absence of fluting upon the shafts, the numismatic evidence is likely to be unreliable, due to the obvious limitations of the medium. However, generally speaking, some attempt is made to conventionalize the forms of the capitals to a motif recognizably Corinthian, Ionic or Tuscan. T. J. Donaldson (op. cit., facing p. xvi) has an amusing plate in his book of some varieties of capitals noted by him among coin types.

Another detail which can seldom be implicitly relied upon in the coin types is the number of steps leading up to the pronaos floor level, if the structure is of the podium type. When it is archaeologically certain that steps existed, it has been found that the coins always indicate them, but their number cannot be accepted as evidence. The Antonine coins showing the temple of Dims Augustus (Pl. VI, 5), for instance, have steps varying in number from one to six, yet all these coins are approximately of the same date and certainly show the same phase of the building. In one instance, however, the surprising accuracy of observation in regard to the arrangement of an entrance terrace and stairway was of great value in demonstrating the identity between two structures shown on two different issues, i.e., Sol Invictus (Pl. VII, 6) of Elagabalus, and Jupiter Ultor (Pl. VIII, 2) of Alexander Severus. 18 On the other hand, inaccurate, or perhaps stylized drawing has created a considerable obstacle to the identification of the representation of the temple of Venus Genetrix (Pl. IV, 4) on certain coins of Trajan. In this latter case, however, the steps may have been added to the front of the podium (whereas in actuality they were at the sides and not visible from the front) in order to avoid implying that the coin type was a putative elevation only, and not a real structure (see p. 11).

As far as the roofs of the various types of structures encountered on the coins are concerned, no reason has ever been found to doubt the reliability of what is shown. In other words, when it is known that a roof was gabled, it appears so on the coins, when flat, flat and when domed, domed.

Decorative friezes and pedimental tympana can scarcely be expected to be minutely reproduced on the coins. As far as the tympana are concerned, it was deemed sufficient to indicate at most a few standing figures and some angle-filling objects. Even this abridgment was increased, and often a wreath or globe is placed in the tympanum to indicate more elaborate works. The Flavian coins of Jupiter Capitolinus (Pl. III, 1, 5) show this convention at work, since both highly circumstantial and highly conventionalized tympana decorations occur. In one instance, the temple of Concord (Pl. II, 1) of Tiberius, there is good reason to believe that the pedimental groups of statuary were moved by the die-cutter from the tympanum to the cornice of the roof for reasons of space.

The acroterial sculptures which appear on the numismatic representations are, generally speaking, reliable as evidence of actuality, but even here different dies show different arrangements for the same building, as can be ascertained from a study of those coins showing, for instance, the Antonine temple of Divus Augustus (Pl. VI, 5). It is probable that these changes are due to varying observations by the die-cutters. However, where care has been taken to show plainly sculptured acroterial figures, it may be assumed that such figures once existed upon the roof of the building in question, although their relative positions may not be exactly preserved. This latter proviso does not apply to the decoration found on the apex of the roof, which as far as can be verified is, when shown at all, a faithful reproduction of reality. Very often what must have been elaborate roof decoration is abbreviated by means of some simple geometric form. There seems to have been the same attitude toward these details as toward the number of columns—abbreviation was allowable, but not addition.

The great elaboration of carved detail usual upon Roman cornices and entablatures is never shown upon the coins. It is abbreviated by means of pearled lines, or by some other simple method. This can be observed especially upon the long series of the temple of Venus and Rome , where all sorts of combinations of plain lines and curved lines and dotted lines are used for this purpose. Antefixes are occasionally carefully drawn, as for instance on the temple of Vesta in Foro (Pl. I, 2) of the gens Cassia, but generally they are omitted, or radically abbreviated.

Incidental decoration of a distinctive nature, such as honorific columns and free-standing statues, is generally shown on types preceding the late second century A.D., as the Concord (Pl. II, 1) of Tiberius or the Venus and Rome (Pl. IV, 5) of Hadrian shows. The Flavian coins showing the temple of Vesta in Foro (Pl. III, 4) are unique sources for the knowledge that this building once had flanking terraces and statues. In various other instances, coin types serve as evidence for the existence of incidental temple furniture, which ordinarily would not otherwise be suspected. After the late second century A.D., such details tend to be suppressed.

Cult statues, although beyond the scope of a purely architectural investigation, need a word here. They are often shown as standing within the colonnade of a temple whose center columns have been given an exaggerated intercolumniation. That these wide intercolumniations are not records of reality can be demonstrated in almost every instance where the façade of a temple or its foundations have been preserved, e.g. the temple of Concord (pl. II, 1) of Tiberius and the temple of Venus Genetrix (Pl. IV, 4) of Trajan.

In summation, we may say that when Roman coin types show temples in Rome, the actual building was used as a model by the die-cutter. Care was taken to record the essential and distinctive features of the building, so that identification would be easy for a contemporary, if not necessarily so for us. Details are often abbreviated or even eliminated, but are never added. It is only in the very late series showing the temple of Venus and Rome , and in those relatively few other instances where temples are used as background for scenes of action, that little reliance can be placed upon the spe- cific evidence offered by them. But even in this latter group, an effort is made to indicate the salient and distinctive features of the building, e.g. the temple of Jupiter Capitolinus (of Domitian's ludi saeculares series, Pl. III, 5), of Vesta in Foro (of Julia Domna, Pl. V, 5), and of Concord (of Alexander Severus and Orbiana, Pl. VIII, 6).

Four main methods of representation can be distinguished among the various types composing the series studied herein. They are: (1) the full-front "simple" type; (2) the full-front "ornate" type; (3) the "vista" type where perspective is used; (4) the "background" type, wherein the temple building serves as the background for a religious or political ceremony. The first category is found throughout the entire period covered by this study. It introduces the architectural temple type in the Volteia coins showing the temple of Jupiter Capitolinus, dated 88 B.C. (Pl. I, 1), and it ends the series with the coins of Alexander Tyrans showing the temple of Venus and Rome , dated 311 A.D. (cf. Pl. IX, 8). Within this period twenty-two structures are shown in the simple frontal style. 19

It will be noticed that this method of representation embraces the majority of temples in Rome which are shown on coins, twenty-two out of forty-three, and is spread rather evenly throughout the period in which architectural representations are found upon coins. In the first century B.C., it is the most common method, since Method 2 appears not at all in that period and as will be developed below only one instance each of the use of Methods 3 and 4 is to be found. Some of the structures listed here will appear again in lists of the other methods, as, for instance, Jupiter Capitolinus (Pl. III, 5) of Domitian (the ludi series), Jupiter Tonans (Pl. III, 6)) of Domitian (the ludi series), Diva Faustina (Pl. V, 4) of Antoninus Pius, and several others. When Method 1 is used, the representation takes up all the available space on the flan, and is characteristically stolid in the Roman portrait sense. In the beginning of the third century A.D., a method of stylization is developed within this category which is never lost. This is a phenomenon observable in other branches of Roman art at the same time, and it is not surprising to find it upon coins too. The best exponent of this trend is the late series of coins showing the temple of Venus and Rome. The stylization of the architecture found here compares favorably with that, for instance, upon the tensa Capitolina of the third century A.D., 20 or better, the Constantinian reliefs from the arch of Constantine. 21 Frontality and a reduction to basic essentials characterize the style.

The temples shown by the second method, with what has been called the frontal "ornate" style, are less nu- merous. 22 These coins cover the period between 35 A.D. and 173 A.D. The criterion for this group is the occurrence of a wealth of detail in the representation of the structures. Chronologically these coins cover the period in which Rome was at her greatest, and it is this circumstance which is reflected in the richness of the dies. All the types in this group were struck to commemorate imperial building activity, with the possible exception of Isis. Two of the types, Concord and Vesta in Palatio (Pl. II, 1, 4), were struck by Tiberius and constitute the only temple types issued by this emperor. Three are Flavian, Isis (Pl. III, 2), Jupiter Capitolinus (Pl. III, 1), and Vesta in Foro (Pl. III, 4), and with the exception of some minor issues of Domitian constitute the only temple types struck by these rulers. There is thus, roughly in the last two thirds of the first century A.D., a small group of coin types which were struck to commemorate imperial religious building activity, and which are isolated by their ornate style. A similarly dated ornate style is found in Roman architectural decoration, e.g., the ornaments of the temples of Concord 23 and of Vespasian, 24 and it is not without interest to see the same attitude influencing the die-cutters when they are called upon to cut designs celebrating the erection of buildings.

The other types of this group are all of the second century A.D., and belong to Trajan, Hadrian, Antoninus Pius, and Marcus Aurelius, that is between 107 A.D. and 173 A.D. Again these coins of Method 2 celebrate imperial building—construction or reconstruction. 25 The principle can be formulated that when temples built by imperial command from about the mid-first century A.D. through the first three quarters of the second century A.D. are shown upon coins, the buildings are shown in frontal style with richly detailed sculptural decoration. There are, however, certain exceptions to this generalization, e.g. Venus Genetrix (Trajan; Pl. IV, 4), Diva Matidia (Hadrian; Pl. V, 3), Bacchus (Antoninus Pius; Pl. VI, 1), all of which, with others, are exponents of Method 3, which should be recognized as an alternative for the principle stated above.

The coins of Method 2 show, then, that an attitude developed in Rome toward detail, both in monumental and minor art, which begins with Tiberius and continues through the Antonine period. This is an attitude of detail for detail's sake and can be seen in the monumental painting of the III, IV, and so-called V Pompeian styles, 26 as well as in architectural decoration, 27 monumental relief sculpture, 28 and freestanding sculpture. 29 This tendency does not hold the field in solitary supremacy, but shares it with that simple style discussed in Method 1 and with the style of Method 3, which, in its desire to give detail, is closely allied to Method 2, adding only an atmospheric quality. Late Antonine monumental art shows the same loss of detail which suppresses the style of Method 2 and results in the supremacy of the style of Method 1 at the same date as upon the coins. 30

Method 3 is exemplified by the representation of fourteen temples. 31 The choice of these has been based upon the demonstrated desire of the engraver to place his structures in space and not to show them merely as façades. The group covers the period between ca. 42 B.C. (Neptune), and the mid-third century A.D. (Sol in Circo Maximo of Philip Sr.). It may well be asked if it is not as usual in occurrence as Method 1, which had a similar chronological spread. However, the number of different buildings in this group is less than in the group of Method 1 (fourteen to twenty-two), and a study of the dates of issuance will show that the coins in Method 3 are concentrated in the second century A.D., with some stragglers both before and after.Therefore the difference between the all-pervasiveness of Method 1 and the concentrated character of Method 3 is clear. This concentration of the use of Method 3 in the second century A.D. illustrates what so often happens in the history of art, i.e., the exceptional popu- larity at a certain period of one particular manifestation of style which has been present before in an unstressed form and which will continue in an unstressed form. Thus those coins and medals which show the use of Method 3 to the best advantage are Bacchus (Pl. VI, 1), Diva Matidia (Pl. V, 3), Divus Julius (Pl. V, 1), Fortuna Redux (Pl. VII, 4), Sol in Circo Maximo (of Trajan; Pl. IV, 2), and Venus Genetrix (Pl. IV, 4), all of the second century A.D. and struck by either Trajan, Hadrian, Antoninus Pius, or Marcus Aurelius. Yet the tendency for a spatial representation is already apparent in the first century B.C. upon the Neptune coins (Pl. I, 5, ca. 42 B.C.). The conception of this early type, however, is not in the grand style of the later second century ones. A likely comparison with this phenomenon upon the coins is to be seen in the construction of the Fora at Rome. The tendency to have a monumental forum is there from the beginning, but its fulfillment as a spatial unit is not seen until the great creation of Trajan. Similarly on the coins, the desire for a monumental spatial composition is observed at an early date, but its achievement is not seen until the second century A.D. Exceptions in this group of Method 3 are the medallions of the third decade of the third century A.D., which show the precinct of Sol Invictus / Jupiter Ultor (Pl. VII, 6; VIII, 2). These are struck some fifty years after the last of the second century items. The lapse in the use of Method 3 and its failure to be used after these late medallions points to a conscious revival of an earlier style being the cause of its sudden and unexpected appearance. It is not possible to base any conclusions of continuity upon the use of Method 3 in the still later contorniates of the fourth century A.D., showing temples of Magna Mater and the temple of Sol in Circo Maximo, since these objects were the results of copying, as the obverses so plainly show.

Unlike Method 2, Method 3 is used equally for structures built by imperial command and for others which were already existing. Of the second century group mentioned on page 27, four commemorate building activity (Sol in Circo Maximo, Venus Genetrix, Diva Matidia, and Bacchus), and two do not (Divus Julius and Fortuna Redux). In the first century A.D. group, i.e. Janus, Jupiter Capitolinus, Jupiter Tonans and Minerva, one (Janus) commemorates a political event and the other three a religious event. The first century B.C. occurrence of the method on the Neptune coin commemorates putative building activity as demonstrated on page 11.

The study of the coins in Method 3 establishes the principle that only in the second century A.D. are ambitious spatial compositions to be found among the architectural coin types of Rome. Method 3 aims at greater realism, and so does Method 2, although different ways are used to achieve the goal. After the second century A.D., both methods are abandoned in favor of a system of schematized frontality which will be discussed below.

A comparison of the development of architectural representations in relief sculpture shows a very similar development to that seen upon the coins of Method 3. A good first century B.C. example of this "perspective" tendency is offered by the panel from the Ara Pacis showing the sacrifice of Aeneas. 32 A building appears in this scene in a manner comparable to the temple of Neptune upon the gens Domitia coin. 33 A Claudian relief in the Villa Medici with a sacrificial scene shows a continuation of the perspective approach. 34 The type of "illusionism" seen in the reliefs upon the arch of Titus furthers this attitude which can be said to culminate, but in a more plastic sense than the Flavian atmospheric one, in the reliefs upon the column of Trajan. The analogy between the coin types and the monumental architectural ensembles of the second century A.D. has already been pointed out. The coin types in Method 3, therefore, are exactly reflecting the trends seen in the major art production of the period in which they appear. A parallel circumstance can be demonstrated for the types of Method 1 which, it will be remembered, continued to flourish throughout this period too.

The coins grouped by the use of Method 4 always show the temple in the background of a civil or religious scene. The temples, therefore, form a kind of back-drop and usually are accorded a sketchy treatment. Fifteen buildings are comprised in this group. 35

Many of these buildings have already appeared under other categories, especially in the group of Method 3. It is obvious that both methods have a rather similar result, since both aim to place the building in space, each, however, using a different means to achieve this end. Since Method 4 always uses figure groups in conjunction with the temple, the isolation of these types appears to be a valid one in contradistinction to those of Method 3.

With the exception of Jupiter Feretrius, Apollo Palatine, Jupiter Capitolinus and Jupiter Tonans, all of these types are dated from the mid-second century A.D. to the mid-third century A.D. There is here, therefore, the same phenomenon as was observed in the case of Method 3, the sudden intensification of a type known before but hitherto neglected. It is also of significance that this method receives its greatest use in a period in which there was no great religious building activity in Rome. The struggle of the religion of the State with various foreign cults, including Christianity, had already begun, and dogma was of more importance than architecture. Accordingly more scenes of ceremony are found, and fewer independent buildings. In almost every instance, nonetheless, an effort has been made by the die-engraver so to depict the structure in the background that it can be identified easily. That he has not been successful in every case is more an indictment of our scholarship than of his ability. Exceptionally good examples of his desire for clarity are to be seen in the Concord, Fortuna Redux, and Vesta in Foro issues. Even the first century B.C. representation of the temple of Jupiter Feretrius is remarkable for its adherence to what we know from the sources about this lost structure. As a general rule, it may be said that temples frequently appear in the background of religious or civil ceremonies on coin types of the second and early third centuries A.D. and only rarely before and afterward. During that century, the temple in the background is usually shown in the full-front method, but exceptions exist, such as Divus Julius of Hadrian and Fortuna Redux of Marcus Aurelius. During the first century A. D. such scenes are rare and when they occur the temple may be full-front (Apollo Palatine) or in perspective (Jupiter Capitolinus, Jupiter Tonans). The single exponent of Method 4 in the first century B.C. has the temple in full-front (Jupiter Feretrius).

Putting aside considerations of composition discussed above in Methods 1–4, a study of the style of the die-cutter's architectural representations will prove interesting. The issues of the first century B.C. 36 are all stylistically similar as far as the treatment of the relief is concerned. The relief is high and clear, and the individual elements of the various structures are carefully distinguished. This sense of clearness and logic in drawing which is found on these coins is never entirely lost dur- ing the period in which architectural coin types occur. Some of the fourth century A.D. contorniates, as for example the one showing Magna Mater (Sabatier, XI, 4), appear to be confused in their drawing, but such examples are very rare and outside of the general stream.

The issues of the Julio-Claudian period generally carry on this clear style, but the effect is more grandiose and more complicated. 37 With the exception of Nero's Vesta series, and his gold Janus coins, which imitate the first century B.C. style, all of these types show an increase in detail and a more realistic sort of observation than was to be seen in the first century B.C. group. There is an attempt to include more of the surroundings in the picture, vide Concord and Vesta in Palatio, and to set the temple in space, vide the bronze Janus coins of Nero, which have an illusionistic tendency. Such a development is absolutely natural when we consider the other branches of Roman art of the period, which are leading toward the illusionism of the Flavian period, as seen for instance on the arch of Titus. The conservatism which generally characterizes numismatic representations of all ages, however, is demonstrated by the retention of Method 1 as well as the style of the first century B.C. in the Nero gold Janus series as contrasted with the more progressive bronze Janus type of the same emperor.

The Flavian issues carry on the scheme of undeveloped illusionism as seen in the Julio-Claudian period. However, the ludi saeculares coins of Domitian, showing Jupiter Capitolinus (Pl. III, 5) and Jupiter Tonans (Pl. III, 6) are strongly illusionistic, and perspective drawing of temples whose rear parts fade off into nothing is common. The same emperor's issues of Minerva (Pl. IV, 1) show a similar type of illusionistic drawing. Along with this advanced style, such semi-conservative ones as those of Isis (Pl. III, 2) of Vespasian, or Vesta in Foro and Jupiter Capitolinus (Pl. III, 1, 4) of the three Flavians, are found together with such thoroughly conservative ones as the Vesta in Palatio (Pl. III, 3) series. In effect the Flavian issues constitute a microcosmos containing the past and future development of pictorial architectural numismatic representation.

The Trajanic types with the exception of Vesta in Foro (B.M.C., III, 22, 22), a restoration of a Republican type, and Divus Nerva (Pl. IV, 3) which is in the frontal style of Methods 1 and 2, are all in the typical plastic illusionistic vista method of the second century A.D. The vista method encountered here compares favorably with that seen upon the column of Trajan. It is important to notice that the architectural coin types are taken over by the illusionistic method of drawing as completely as are the more monumental arts of the period.

With Hadrian, in keeping with his classical revival, the illusionistic style is abandoned for the only impor- tant architectural type, the temple of Venus and Rome (Pl. IV, 5), although the actual setting of the temple of Venus and Rome would have lent itself admirably to the illusionistic method, since the temple is placed in an area between great porticoes (cf. the temple of Diva Matidia). It is retained, however, in the minor issue showing the temple of Divus Julius (Pl. IV, 1).

This opposition between the two main methods of architectural representation, frontal or perspective, classical or illusionistic, exemplifies the struggle in the other fields of Hadrianic art between the "Greek" and "Roman" styles. In monumental art this struggle can be seen in the contrast between the neo-classical frescoes of the tomb of the Nasonii 38 and the completely illusionistic fresco from the Villa of the Quintili. 39

The Antonine issues demonstrate the increasing abandonment of illusionism and the increased use of the stylized frontal type. Only one issue, Bacchus (Pl. VI, 1) of Antoninus Pius, is in the illusionistic style, the others are all in the frontal style, although sometimes in the figured style of Method 4, which may be construed as illusionistic to a certain limited extent.

An unexpected return to the illusionism of the early second century A.D. is found in the coins and medallions showing the temple of Sol Invictus / Jupiter Ultor (Pl. VII, 6; VIII, 2) struck by Elagabalus and Alexan- der Severus. As noted above, this circumstance seems to be the result of conscious copying of earlier material. The dominant frontal quality of the late Antonine period can be seen working even in these "perspective" views where the emphasis is on the frontal plane of each succeeding layer of the composition, rather than upon the creation of a sense of depth. That this quality of frontality is not restricted to the coins can be observed by comparing the reliefs of the column of Marcus Aurelius 40 with those of Trajan's column.

It is this frontality coupled with schematization that finally conquers the entire field of numismatic architectural representation. All the third and fourth century A.D. issues, with the early third century exceptions noted in the paragraph above, are in this style. No attempt is made at realism, but symbolism is stressed. It is due to this development that the frequent representations of the temple of Venus and Rome (Pl. VIII, 4, 5; IX, 1, 3, 5–8), the symbol of Empire, are found. In this style it is no longer necessary to be faithful to details upon the actual building, but rather to an idea expressed by the representation of that building. As a result are found the confusing varieties showing this temple with from two to eight columns and with the sculptured detail schematized in a hundred different ways. However, the idea behind the representation is always the same, namely, to symbolize the Empire. That adherence to reality was not entirely abandoned is shown by the everpresent seated statue of Roma within the colonnade and by the always rectangular shape of the structure.

The numismatic architectural representations of temples in Rome, as this survey of their stylistic development establishes, follow the general stylistic developments seen in Roman monumental art. Lags such as are found in the use of the fully developed illusionistic style can with good reason be attributed to the conservatism normally found in coinage, although the ludi series of Domitian demonstrates the possibility of a rapid, if not persistent, adoption of a new style by the die-engravers.

THE chronological list following embodies Roman coins which bear representations of temples of the city of Rome. Plate references are given for those pieces illustrated, with indication of the metal and of the collection from which the coins illustrated derive. For convenience of reference, coins not illustrated herein have citations in italics to published reproductions of their respective types. Following are abbreviations used in the list:

• a.n.s.=American Numismatic Society.
• bmc.=Coins of the Roman Empire in the British Museum, Vols. I–III, 1923–36.
• bmcr.=Coins of the Roman Republic in the British Museum, Vols. I–III, 1910.
• c.=H. Cohen, Description Historique des monnaies frappées sous l'empire romain, ed.2, Vols. I–VIII, Paris, 1880–92.
• gn.=F. Gnecchi, I Medaglioni Romani, Vols. I–III, Milan, 1912.
• ms.=Mattingly and Sydenham, The Roman Imperial Coinage, Vols. I–V, 1923–38. In progress.
• nz.=Numismatische Zeitschrift.
• nnm.=Numismatic Notes and Monographs.

Key to Footnote Indications employed in List:

• §Not able to locate a specimen.
• ¶ Temples not specifically identified, but grouped together as VOTA because their inscriptions concern the annual vows.
• *, †, ‡=Published in British Museum Catalogue under one of these conventional signs instead of a number.

Contorniates of the fourth century show Janus, Magna Mater, Sol in Circo Maximo, and Venus and Rome. Since these copy earlier coins and medallions, they will not be discussed here.

A CERTAIN number of coin types with architectural representations which have hitherto been considered to show temples in the city of Rome have, after examination, been rejected as either not showing a temple but a secular building, or as not showing a temple in the city itself. The reasons for this rejection will appear in a forthcoming study. The types which have been rejected are listed below with a reference where a typical piece may be found. (For abbreviations, see Appendix A, p. 39.)