CHAPTER TWO

DROGO’S COPY OF THE

Handbook of 809

A Simulacrum of Celestial Order

The celestial host surrounding medieval Metz maintained an otherworldly permanency. On Bede’s own authority, the realm of the fixed stars and the planetary wanderers would survive the Apocalypse, bathed in the radiant splendor of the Almighty, who purged the unclean with fire. In chapter 70 of De temporum ratione, Bede had informed Drogo and all learned Franks that prior to the donning of “everlasting bodies,” “the airy heaven will shrivel up in fire, [but the heaven] of the stars will remain undamaged. In fact, the heavenly bodies will be darkened, not by being drained of their light, but by the force of a greater light at the coming of the Supreme Judge.”¹ The texts and images of the fixed stars in a book such as Drogo’s copy of the Handbook of 809 were signs of stability in a world of transience. Even the hope and promise of the Apocalypse, which played a fundamental role in the evangelistic mission and doctrine of renewal within the early medieval church, could not undermine these integral components of God’s observable creation. Time was consistently restored through the annual and daily cycles of the heavenly bodies, regulated by the sun and moon.²

For the Frankish princes who owned copies, the combined moments of meditative reflection upon the texts and star pictures of the Handbook of 809 were part of the unfolding salvation history of the Franks. Individual episcopal study by a prelate such as Drogo indirectly helped fulfill the soteriological mandates of the early medieval Christian church that called for sacred study of the liberal art of astronomy.³ Drogo’s personal copy of the Handbook of 809 was thus an ephemeral portal—bound for destruction within the Apocalyptic flames of penitence—that nevertheless opened his mind and spirit to the eternal mysteries of the heavens.

Paul Edward Dutton has argued in this context that courtly encomia likening worthy bishops to stellar vicars of Christ had more than a literary significance, operating as a topos for a poet such as Sedulius Scottus (fl. ca. 840–60) within the Frankish realm. Frankish kings such as Charlemagne or Louis the Pious achieved a veritable “Christianized Carolingian apotheosis” that could be extended by royal blood to their episcopal sons, nephews, and cousins, legitimate or otherwise, in service to the church.⁴ As Dutton notes, Sedulius, apparently an itinerant Irish monk, celebrated Bishop Hartgar (served 840–55), under whom he had served in Liège, perhaps achieving the rank of scholasticus (roughly equivalent to a headmaster) within Hartgar’s cathedral school dedicated to Saint Lambert.⁵ It is highly likely that Sedulius also composed a number of lyrical panegyrics in Metz for Bishop Adventius (served 858–75).⁶ Sedulius specifically makes Hartgar “earth’s most radiant star,” handpicked by Aurora in one poem,⁷ and in another compares the physical ascent of a tower to the expansion of the prelate’s mind, making the fulfillment of his episcopal and homiletic duties to the people of Liège possible:

The daughter of Zion rejoices in such a shepherd, and the rich and poor exult with joy.

He builds a lofty tower, a hundred cubits high, so that he may ascend above the stars.

Climbing the stairway rising towards heaven, he instructs his flock with sage words and examples.⁸

Drogo, similarly preparing himself through study of his personal copy of the Handbook of 809, thereby “wards off wolves and rescues his lambs.”⁹ It is within this framework, joining personal spiritual study to the establishment of a celestial Carolingian hagiography, that the text-image relationship in the Handbook of 809 must be addressed, and Drogo’s personal book is the best example of an official spiritual-yet-scientific manual disseminated throughout the Frankish realms for such use. Stephen McCluskey has said of such “astronomical and computistical anthologies,” otherwise known as “star catalogues,” that “their goal was not astronomical observation but artistic and mythological edification.”¹⁰ More important, study of the stars brought Carolingian bishops into the heavenly realms, where the planets and fixed stars served as stepping stones to intellectual freedom and spiritual enlightenment, purifying the hearts of these prelates while simultaneously preparing them to accomplish their spiritual mission on earth by bringing their fellow Franks along with them.¹¹

In this chapter, the central theme of part 1 of the present volume continues to unfold: the art-historical significance of Drogo’s personal copy of the Handbook of 809 requires an investigation into the iconographic history of classically inspired but notably reinterpreted images of the heavens. This chapter focuses primarily on a detailed review of the pictorial program of constellations in Drogo’s handbook. After an initial explanation of Aratean traditions, I offer a complete iconographic analysis of the Greco-Roman traditions that informed the symbolic properties of the miniatures in Drogo’s handbook. In addition, specific attention is given to images such as Chiron the Centaur in order to supply a cogent example of Carolingian exegetical emendation and strongly advocate for Carolingian painterly creativity, as argued in the introduction. As seen in chapter 1, a thorough examination of Drogo’s copy of the Handbook of 809 permits a celebration of the ways in which students of astronomy and computus created a dynamic record of their spiritual science.

Here in chapter 2, the profound appreciation for classical iconographic traditions manifested by the Carolingian prelates and painters who created Drogo’s copy of the Handbook of 809 is contextualized, in terms of both the mythological narratives informing image types and the scientific soteriological benefits of studying the star pictures along with their corresponding texts. In every viable instance, Frankish painters demonstrated an appreciation of the traditions they transformed, creating a novel Christian presentation of the heavens one constellation at a time. At the end of this review, some closing remarks concerning Panofskian iconographic traditions build upon the comments in the introduction and conclude part 1.

History of the Aratean Pictorial Cycles and Their Texts

This novel spiritual justification for the Carolingian star catalogues supplied the impetus for the original description of the constellations offered to readers of the De ordine, which was systematically derived from an admixture of mythological lore and the presentation of the constellations modeled after star globes. The textual and visual considerations of vital significance for the De ordine hark back to Aratus of Soli, who came to Pella in 276 B.C.E. and composed the Phaenomena at the Macedonian court of Antigonus Gonatas, ca. 276–274 B.C.E.¹² According to a somewhat apocryphal legend, Aratus did not derive his poetic description of the constellations, their stellar utility for discerning temporal changes, or their meteorological significance directly from the heavens.¹³ According to Cicero, Aratus was a well-trained literary dilettante from Athens without professional astronomical skills: “Eudoxus of Cnidos, who was a pupil of Plato’s . . . marked on the globe the stars that are fixed in the sky. Many years after Eudoxus, Aratus adopted from him the entire detailed arrangement of the globe and described it in verse, not displaying any knowledge of astronomy but showing considerable poetical skill.”¹⁴

Douglas Kidd has valiantly defended Aratus against such Ciceronian naysayers, arguing persuasively that Aratus neither imported wholesale nor uncritically endorsed Eudoxus’s interpretation of Greek astronomy. On the contrary, Aratus fundamentally altered the perspective from which he drafted his 1,154 lines of polished verse, adopting the standpoint of the stargazer and earthbound student of the heavenly host.¹⁵ This radical move reflects more than authorial repositioning. The novel voice of Aratus harnessed the heavens as a cosmological realm populated by mythopoetic heroes and celestial wonders replete with signposts for knowledgeable stargazers to discern. The abundance promised to farmers who augured well the agrarian signs, and the safe journeys vouchsafed to sailors who navigated auspicious times for travel, had been established by none other than attentive Zeus. The stars and their constellations were a road map to Greek or Frankish survival, and arguably even to individual success, according to Aratus’s proem:

Filled with Zeus are all highways and all meeting-places of people, filled are the sea and harbours; in all circumstances we are all dependent on Zeus. For we are also his children, and he benignly gives helpful signs to men, and rouses people to work, reminding them of their livelihood, tells when the soil is best for oxen and mattocks, and tells when the seasons are right both for planting trees and for sowing every kind of seed. For it was Zeus himself who fixed the signs in the sky, making them into distinct constellations, and organized stars for the year to give the most clearly defined signs of the seasonal round to men, so that everything may grow without fail.¹⁶

It is interesting that the providential role of divine oversight in the opening verses of Aratus’s Greco-Macedonian Phaenomena (or proem doubling as a panegyric to Zeus) also provides a pagan foil to a kindred idea of Christ the Good Shepherd that was appropriated by the early Christian church. The later Latin derivatives of the Aratean textual traditions, culminating with the De ordine in the Handbook of 809, attest to this development, which further enhanced the individualistic turn in Aratus’s transformation of the earlier information about the stars reported by Eudoxus, his fourth-century predecessor (fl. 368–365 B.C.E.). Other notable deviations from the lost text of Eudoxus include Aratus’s effort to align the advent of the constellations appropriate to seasonal zodiac signs with their respective positions on the solstitial or equinoctial colures.¹⁷

In part 2 below, chapter 3 examines the stylistic concerns of importance for a revisionist history of Carolingian painting in Metz, while chapter 4 details the significant transformations of long-standing Aratean traditions in the service of a Carolingian scientific soteriology. (Students of philology rather than art history may prefer to read chapter 4 before returning here, recognizing that an alternate trajectory through the text will affect the presentation of ideas and the overarching argument of this book: namely, that Drogo’s copy of the Handbook of 809 documents in its texts and imagery a comprehensive Carolingian effort to preserve a simulacrum of celestial order.) In the present chapter, a survey of the star pictures in Drogo’s copy of the Handbook of 809 permits an exhaustive iconographic review of the origins and nature of the forms of the zodiac and constellations that were available to ninth-century painters.

Zeus’s array of constellations, which resulted in part from the legendary catasterisms of Greek mythology, were stolen in Promethean fashion from the pagan deity and set within the minds and hearts of Frankish students of the liberal art of astronomy. The De ordine recounts in general terms the mythic heritage of the constellations and thereby announces from the outset the twin themes of the text:

Here is the order and the placement of the constellations, which were fixed in the heavens by the grouping of several stars into signs. The kinds of forms are believed to have been admitted into the heavens from some models or fables. The arrangement of the natural forms did not yield their names, but human conviction gave the numbers and names to the stars. But, in accordance with Aratus the number of stars belonging to each sign has been appointed; the description should follow according to the order, which is his own.¹⁸

Not Zeus but the pagan ancestors of eighth- and ninth-century Christians thus clustered the stars and then gave these arbitrary groupings the names by which they remain known to this day. Advancing a personal outlook, the Frankish author of the De ordine relied upon the same powers of imagination and intellect that earlier people had used to designate the constellations. In so doing, the author of the De ordine followed Aratus’s lead, but in keeping with standard interpretations of the Frankish renewal: the individual early medieval Christian’s position within the cosmos (and the concomitant placement of his microcosmic soul) required a novel celebration of Aratean tradition. The Franks reoriented their understanding of the heavens and pointed Carolingian astronomical study in a joint scholarly and spiritual direction, one that permitted revision of classical narratives that had come down via lines of textual transmission discussed here and in chapter 4. The emphasis on the individual person’s observation of the heavens, inherited from Aratus and coupled with a sincere desire to reinterpret tradition to fit the needs of a new ruler and polity, were therefore parlayed into an unexpected Christocentric, soteriological focus as a driving motivation for the Carolingian study of astronomy and the stars. As explained in the introduction and argued in this chapter, the images of the constellations themselves attest to the scientific sacralization of antiquity and make possible the Frankish scientific soteriology witnessed by novel presentations of the constellations.

Given these preliminary considerations, it is not surprising that a learned translator of astronomical texts from Corbie adopted and adapted the order of the star pictures belonging to Aratean textual traditions for his Christian presentation of the constellations. This supplied a meaningful structure for the De ordine, transported by Adalhard of Corbie to the Aachen synod of 809. In all Aratean texts, the standard way to adduce the constellations, including the zodiac, was to begin with the celestial hemisphere north of the ecliptic and continue with those constellations found south of the ecliptic.¹⁹ Hence the depictions of the constellations in Drogo’s copy of the Handbook of 809 will be described in the order in which they appear, adhering to and revealing the standard text of the De ordine. This ordered presentation of the constellations recurs in other texts derived from the Aratean Phaenomena.²⁰ Since there is such a high degree of standardization in the sequence of the cycle, the idiosyncrasies in discrete pictorial programs and alterations associated with various textual recensions become significant. Before returning to a discussion of Adalhard, it is useful to review precisely how he and others envisioned the structure of the celestial sphere.

The ecliptic is the oblique that circles the celestial sphere on which are permanently situated the fixed stars, composing the constellations. The entire celestial sphere rotates daily in a clockwise direction about a stationary earth. This line receives its name because it defines the path of the sun throughout the course of the year. Since all planets—including the moon and sun, according to the Ptolemaic model of the solar system—revolve about the earth in a counterclockwise direction, the only place the sun, moon, and earth can overlap is on the ecliptic, at the center of the geometrically distributed twelve-part zodiac belt, causing eclipses.²¹

According to a Carolingian understanding of Plinian and classical astronomy, the zodiac belt extends 6 degrees north and south of the ecliptic, defining a 12-degree diagonal swath of space through which the planets move latitudinally, according to their inclination, relative to the ecliptic. The equator defines the central circumference of the earth and is complemented by the so-called celestial equator exactly parallel to it.²² The parallel tropics of Cancer to the north and Capricorn to the south of the equator identify zones of roughly 23.5 degrees of geographic latitudinal space, respectively. The oblique ecliptic intersects the tropic of Capricorn at its southernmost point in the region once occupied by the constellation Capricorn, whereas it intersects the tropic of Cancer in the former region of the zodiac constellation Cancer at its northernmost point. This also explains the origin of the tropics’ names.²³ Because of the phenomenon of precession, or the perpetual shift of all constellations on the celestial sphere from west to east (i.e., counterclockwise) over time, the geometric 30-degree sections of the ecliptic defined as the signs of the zodiac do not align perfectly with the constellations that bear their names and through which the sun passes annually. In other words, stars from the constellation Pisces are actually located within the spatial zone identified by the geometric zodiac sign of Aries; Aries, the constellation, has entered Taurus’s territory; and so forth.²⁴

Again, according to the geocentric model of the universe, the fixed and wandering stars (planets) rotate around the earth. There are two conflicting motions, which typically cause great confusion for those trying to understand a medieval view of the heavens. On the one hand, the sun and all planets make their way through the zodiac signs along a counterclockwise trajectory (ignoring for the moment the phenomenon of apparent planetary retrograde motion discussed in chapter 3) that follows the order on the diagram “The Apsides of the Planetary Orbits Within the Zodiac” in Drogo’s copy of the Handbook of 809 (fol. 65v; see fig. 5). Thus we see Aries at 9:00, designating the beginning of solar springtime renewal, followed by Taurus, Gemini, Cancer, Leo, Virgo, Libra, Scorpio, Sagittarius, Capricorn, Aquarius, Pisces. Simultaneously, every heavenly body within the geocentric cosmological model, including a fortiori the celestial sphere of the fixed stars, rotates about the stable earth once daily in an apparently alternate clockwise motion.²⁵ These two aspects of the medieval celestial model derived from Ptolemy permit the daily passage of the sun and moon as well as the seasonal shifts linked to the liturgical calendars, regulating lives on earth through their rapport with the celestial bodies surrounding them on all sides.²⁶ In the words of Aratus, “The numerous stars, scattered in different directions, sweep all alike across the sky every day continuously for ever. The axis, however, does not move even slightly from its place, but just stays for ever fixed, holds the earth in the centre evenly balanced, and rotates the sky itself.”²⁷

Hubert Le Bourdellès has emphasized the significance of Adalhard of Corbie’s involvement in the synod of 809. Adalhard was, as noted in the introduction, named in the report that followed the convocation of the computistical subgroup that assessed the state of the knowledge of computus during the synod of 809.²⁸ It is highly likely that Adalhard brought two vital records of Frankish interest in astronomy with him to the Aachen synod. Those two new Carolingian creations from Corbie became the key illustrated components of book V in the Handbook of 809. Le Bourdellès has fully identified the textual sources for these two records, and this information is key to the story of the creation of Drogo’s copy of the Handbook of 809 undertaken here. The Excerptum (text V.1) supplied an abstract, as it were, of the version of the Aratus Latinus known as the Revised Aratus. The Aratus Latinus was a translation into Latin from the ancient Greek poem described here, the Phaenomena of Aratus. The earlier translation probably took place at Corbie around 735 and then provided the foundation for the Revised Aratus, which is also likely to have been made in Corbie, around 790.²⁹ The most important example of the Revised Aratus is the third-generation copy from Corbie presently located in Paris (Paris, Bibliothèque nationale de France, MS lat. 12957; hereafter the Paris Aratus). Created in the ninth century (probably ca. 810), this copy of the Revised Aratus includes a set of miniatures (fig. 35). The line drawings suggest to Le Bourdellès that both the original copy of the Aratus Latinus translation from Greek into Latin and the lost original copy of the Revised Aratus itself were illustrated with Aratean star pictures.³⁰

These are contentious points, although Le Bourdellès is probably accurate in his assessment. An explanation of the reasons he is apt to be correct is essential for an understanding of the origin of the star pictures found in the Handbook of 809, and a fortiori for those in Drogo’s copy. In addition, for this art history of Carolingian astronomical manuscripts, it is equally important to assess whether lost, hypothetical cycles of star pictures included drawings alone or miniatures as well. It is highly likely that there was an originally illustrated copy of the Revised Aratus made in Corbie. The evidence for an originally illustrated Aratus Latinus is sketchier. There were no illustrations in the early copy of the type A version of the astronomical encyclopedia (according to Borst’s nomenclature) discussed in chapter 1, the Verona compilation.³¹ In the late eighth century, medieval scribes therefore did not consider astronomical compilations in need of artistic illustration. In fact, some later copies of the Aratus Latinus lack illustrations (e.g., Paris, Bibliothèque nationale de France, Ms lat. 7886, ca. 850–75; Brussels, MS 10698, twelfth century). But other copies of the Aratus Latinus were probably illustrated (see fig. 7) with detailed line drawings depicting the stars, like those accompanying the portion of the Basel Scholia devoted to the Germanicus translation with scholia.³²

FIGURE 35

Celestial globe. Paris, Bibliothèque nationale de France, MS lat. 12957, fol. 63v. Photo: BnF.

In any case, Aratean manuscripts were not always illustrated. In the case of the Handbook of 809, however, miniatures constituted an integral component of the early luxury copies of the handbook. The ninth-century pictorial cycle in Paris could have been a later addition to the lost original copy of the text, although this is unlikely. The introduction of pictures into the Revised Aratus could have taken place during the manufacture of the hypothetical intermediary copy, postulated by Le Bourdellès, that probably supplied the immediate precursor for the copy from Paris and another presently in Cologne (Cologne, Dombibliothek, MS 83-II, hereafter the Cologne Aratus; fig. 8).³³

The Cologne Aratus includes a date from the incarnation of the Lord of 798 on folio 14v and was probably completed on art-historical grounds in 805. Hildebald, who was archbishop of Cologne (before 787–818) and archchaplain to Charlemagne, compiled the manuscript.³⁴ Charlemagne decreed that clerics should undertake the mastery of the computus with the Capitulary of Diedenhofen (Thionville) in 805. The coincidence between the completion date of work on the Cologne manuscript (805) and the capitulary has led Anton von Euw to hypothesize a direct connection between Charlemagne’s desire for an astronomical-computistical-pedagogical handbook and Cologne.³⁵ In other words, the Cologne copy of the Revised Aratus was an early attempt to create an astronomical and computistical teaching manual before the Handbook of 809.

Two observations about the Cologne Aratus are noteworthy for the history of star pictures during the Carolingian era. First, the completed miniatures in Cologne, finished by 805, are truly painterly, such as the image of Hercules and the serpent on folio 156v (fig. 8). By comparison, the drawings in the Paris copy of the Revised Aratus supply linear contours (fig. 35). A parallel disparity exists between the paintings in Drogo’s copy of the Handbook of 809 (plate 2) and the linear forms belonging to a set of suitably related copies of the Handbook of 809 that justify their assignment to a specific recension: Monza f–9/176 (figs. 36–37) and Vat. lat. 645 (fig. 38). Painted miniatures were actually the exception rather than the rule among the Carolingian manuscripts with star pictures. The inclusion of painted miniatures, as opposed to linear contour drawings, was clearly intended to convey the superiority and excellence of a de luxe manuscript such as Drogo’s princely copy. Astronomical-computistical compilations of great refinement, replete with painterly miniatures, fulfilled more than a pedagogical purpose. They were also expressions of a patron’s power and influence. Given this conclusion, there is little doubt that the original copy of the handbook compiled in Aachen under the direction of Adalhard of Corbie also included a lavishly painted cycle of miniatures. It was made for Charlemagne.

Second, the Cologne and Madrid manuscripts were both made for patrons who were at one time archchaplains of the Carolingian kingdom, as discussed briefly in chapter 1. Hildebald of Cologne ascended to the archchaplaincy after the death of Angilram of Metz in 791.³⁶ Bishop Drogo served both Louis the Pious (after 834) and Lothar I (after 842) as archchaplain, too.³⁷ Since Drogo’s copy of the Handbook of 809 was made around 830, it is reasonable to see in the painterly production of the miniatures an effort to surpass the quality of painting in Hildebald’s proposal for a handbook.

In any case, the pictorial cycles of the Paris Aratus³⁸ (fig. 35) and Cologne Aratus (fig. 8) are similar enough to warrant the suggestion that their predecessor influenced the order and the arrangement of the pictorial cycle in these manuscripts. The pictorial program in these two manuscripts is also relatively similar to the various copies of the Revised Aratus belonging to the second branch of the bifurcated stemma provided by Le Bourdellès. This further supports his contention that the original copy of the Revised Aratus was indeed illustrated. For example, one of the copies from the other branch of Revised Aratus manuscripts is Saint-Gall, Codex 902 of the mid-ninth century (Stiftsbibliothek Sankt-Gallen, Cod. 902, hereafter Saint-Gall 902; figs. 39–40). Saint-Gall 902 includes the illustrated portion of the text referred to as the Recensio interpolata in earlier literature.³⁹ The image of Aries on page 90 (fig. 39) resembles the basic linear form of the painted image in Drogo’s copy of the Handbook of 809 (plate 9), but in reverse.

FIGURE 36

Gemini and Cancer. Monza, Biblioteca Capitolare, f-9/176, fol. 63v. Photo © Biblioteca Capitolare del Duomo di Monza.

FIGURE 37

Sagittarius, Aquila, and Delphinus. Monza, Biblioteca Capitolare, f-9/176, fol. 67r. Photo © Biblioteca Capitolare del Duomo di Monza.

FIGURE 38

Gemini and Cancer. Vatican City, Biblioteca Apostolica Vaticana, Vat. lat. 645, fol. 58v. Photo © 2017 Biblioteca Apostolica Vaticana. Reproduced by permission of Biblioteca Apostolica Vaticana, with all rights reserved.

This iconographic similarity with manifest variety is a standard phenomenon found throughout the individual Carolingian manuscripts containing star pictures. The similarities derive from the long-standing Aratean precursors under discussion here, but the individual differences in specific codices permit additional meanings beyond straightforward classical references. These novel insights are the true legacy of Carolingian book painters to pictorial art. That does not compromise the significance of the Aratean texts transmitted for posterity. On the contrary, along two intrinsically intertwined trajectories, the value of Carolingian scribes to the accurate preservation of texts can be celebrated equally with the interpretative efforts of Frankish painters who built visually on the work of their predecessors and turned tradition into new ninth-century erudition, replete with the awe and wonder the heavens can induce.

The uniformities displayed by depictions of the constellations in the Revised Aratus manuscripts supply the best reason to argue backward for an originally and similarly illustrated Aratus Latinus. For example, the image of Arcturus Maior (The Great Bear) on page 82 of Saint-Gall 902 resembles the basic linear form of the related images in the Paris Aratus (fol. 64r) and Cologne Aratus (fol. 155r). Moreover, in all three cases, the quadruped faces to the left. Naturally, the general forms of the zodiac and constellations can be traced back to Hellenistic precursors. According to Mechthild Haffner, the finalization of the relatively standard pictorial program of the illustrated copies of the Germanicus translation of the Phaenomena, which influenced the order and form of the pictures in the Aratus Latinus and the Revised Aratus, took place in the late third century. These images were attached originally to a hypothetical Greek copy of the Phaenomena with scholia, referred to in the scholarly literature as Φ.

Haffner has noted that certain miniatures in Aratean manuscripts can only be explained by the scholia introduced into the textual redactions. This suggests that the iconography of the star pictures changed over time or that the pictures were added to the text after its composition (at a potentially much later date). For example, a kneeling Hercules only makes sense in the context of the conflict between the hero and the serpent, described in scholia, which later influenced the creation of the Basel Scholia supplementing the Germanicus translation into Latin of the Aratus poem (fig. 7). In other words, Haffner believed that the definitive set of illustrations that influenced medieval astronomical and astrological cycles of pictures was added to the Aratus tradition roughly five centuries after the composition of the original poem by Aratus, and the textual analyses of Le Bourdellès have corroborated this result.

Under this interpretation, the advent of the codex roughly coincided with the creation of a series of illustrations for the Aratean manuscripts—a series that would later become comprehensive. In the third century, the images adhered to a pictorial logic appropriate for a rotulus in Alexandria. During the transition period from the rotulus to the codex, Haffner has argued, there were linear star pictures without frames accompanying a Greek edition of the Phaenomena with scholia. Interestingly, many of the putative earlier examples of the constellations tended to display the dorsal sides of the constellations, as though they were viewed outside the ring of fixed stars on an ideal celestial globe. Haffner’s ingenious solution to the problem of the transmission of the Aratean text and star pictures was to postulate a new Latin translation of the Phaenomena in the late third or early fourth century. This edition relied upon the Germanicus translation for a model, while the translators also addressed the Greek scholia, thereby creating the accompanying Latin scholia, namely, the Basel Scholia. The miniatures accompanying the copies of the Germanicus translation from around that period onward supposedly displayed true codex illustrations, with blue fields and red frames (fig. 21), recalling the pictures in the Carolingian ninth-century Leiden Aratea.⁴⁰

Colin Roberts and T. C. Skeat have identified the late third and early fourth century as the point at which bound books generally replaced rotuli as the vehicle of choice for the transfer of written information.⁴¹ This finding buttresses Haffner’s hypothesis. Here, however, in assessing the origins of the astronomical miniatures in Drogo’s copy of the Handbook of 809, we must recognize the plausibility of an originally illustrated Revised Aratus. That pictorial cycle probably harked back to an original program created in the third century in Alexandria for the Greek edition of the Aratea with scholia. That edition, as Le Bourdellès proposed, probably provided the model for the pictures introduced into the Aratus Latinus and then into the Revised Aratus, recognizing that with every copy there were opportunities for creative aesthetic decisions.⁴²

FIGURE 39

Aries, Deltoton, and Pisces (Aquilonalis and Australis). Stiftsbibliothek St. Gallen, Cod. Sang. 902, page 90. Photo: Stiftsbibliothek St. Gallen.

Like Drogo’s copy of the Handbook of 809, the Cologne Aratus included miniatures in open fields between single-column text blocks. These de luxe painted astronomical manuscripts preferred to return to the more ancient graphic style of unframed images, which was actually associated with a rotulus. On the one hand, these aspects of the pictorial cycles in early medieval illustrated astronomical manuscripts are emblematic of beliefs about early manuscript illustration espoused by Kurt Weitzmann.⁴³ On the other hand, the pictures in these two codices have received uncommon emphasis. In the Paris Aratus (fig. 35) and Saint-Gall 902 (fig. 40), the linear renderings of the constellations have been squeezed into the available allotted spaces, pushing up against the margins. The linear forms seem to be inconsequential diversions from the dominant text blocks in those manuscripts. Yet not only did the de luxe copies of ninth-century astronomical treatises receive painted miniatures, but those miniatures were also placed in free-standing, unframed open spaces, highlighting the independence and artistic refinement of the forms. It is important to note here that the highly developed tinted drawings of the sister manuscript to the Saint-Gall 902 (fig. 40), the Saint-Gall Cod. 250 (Stiftsbibliothek Sankt-Gallen), were also set apart, according to the design strategy for an exceptionally refined cycle of miniatures. The monastic draughtsmen inserted their highly finished drawings into Saint-Gall 250 in open spaces between single text columns in the last quarter of the ninth century, following the procedure for a more luxurious copy of an illustrated astronomical text.⁴⁴ The one-column format was in no way reserved for the painted manuscripts alone, although the presence of paintings and a single-column format in the Madrid and Cologne manuscripts clearly conveys their exceptional refinement and splendor. This actually opposes one tenet of Weitzmann’s history of early manuscript illustration. Weitzmann has argued for a steady progression from early cycles of illustration in scrolls to the emancipation of gilded images, which resulted in framed manuscript illuminations.⁴⁵ Notably, early medieval artists could accentuate the value of their miniatures without the use of frames.

In summary, it is highly likely that a pictorial program enlivened the original edition of the Revised Aratus, and that this pictorial cycle had a significant impact on the development of future star pictures. The Excerptum from the Handbook of 809 (V.1) was not illustrated, but it was the prelude to the text that included the illustrations, namely, the De ordine (V.2). For this reason, it is highly significant that the Excerptum was derived from the Revised Aratus, as argued by Le Bourdellès, and that the Revised Aratus originally included a set of star pictures like those in the Cologne Aratus.

Le Bourdellès also identified the second text from book V of the Handbook of 809 as an original contribution, which Adalhard brought from Corbie to the synod meeting in 809. The text provided a summary of the Basel Scholia.⁴⁶ The latter name derived from the scholia included within a manuscript likewise containing the fragmentary copy of the Aratus Latinus (and not Revised Aratus) attached to an edition of the Germanicus translation in Basel, as mentioned above. It is known that the scholia in Latin could not have been added to any putative version of the Aratus poem that arose before the imperial period of classical Rome, since the scholia included references to the writings of Publius Nigidius Figulus (active ca. 98–45 B.C.E.).⁴⁷ The Latin scholia were translated ca. 300 C.E. from the commentaries to the lost Alexandrian redaction of the Phaenomena by Aratus, as discussed already. Le Bourdellès followed J. Martin when he argued that the hypothetical Alexandrian text combined the Aratus poem with an introductory text summarizing the ideas of Eratosthenes (ca. 276–194 B.C.E.) on terrestrial or celestial globes. In any case, the commentaries had been added to the Germanicus translation by the fourth century, as Le Bourdellès noted, because the Christian exegete from Trier, Lactantius (240–ca. 320 C.E.), alluded to the commentary in his Divinae Institutiones (I.11).⁴⁸ All of Le Bourdellès’ findings, therefore, corroborate the positions advanced by Haffner.

FIGURE 40

Lyra, Aquarius, and Cygnus. Stiftsbibliothek St. Gallen, Cod. Sang. 902, page 92. Photo: Stiftsbibliothek St. Gallen.

In the third-century Alexandrian corpus, the Aratean descriptions of the constellations also included mythological tales about the constellations and star pictures. These supplemental mythological texts provided an account of the fictional reasons for which each of the constellations appeared in the heavens as the handiwork of the gods. Whenever a person, such as Hercules/Engonasin, or a creature, such as Cancer the Crab, was magically preserved for posterity by the introduction of a constellation in its form and honor, a catasterism had taken place. These catasterisms, described in a text derived from Eratosthenes, also catalogued the stars visible within a depiction of a constellation. The Latin translation of the Catasterisms from the Alexandrian compilation provided a supplemental list of the stars in the heavens added to the Germanicus translation of the Aratean Phaenomena. These insertions into the Germanicus text adduced mythological references and a list of stars, one constellation at a time. This complement constituted most precisely the so-called Scholia Basileensia.⁴⁹ The descriptions formed the basis of the new Carolingian catalogue of the stars within the constellations, the De ordine, albeit without the mythological references.

In 1898, Georg Thiele had already recognized the link between the group of manuscripts containing the De ordine and the Catasterisms in the most important early contribution to the study of medieval astronomical illustration, the Antike Himmelsbilder. Thiele labeled this set of manuscripts the “Phillippicus Class,” because there is another Carolingian copy of the De ordine closely related to Drogo’s copy of the Handbook of 809 in Berlin (Preußischer Kulturbesitz, Staatsbibliothek zu Berlin, MS Phill. 1832; hereafter the Berlin Ordine). Thiele linked, for example, some of the pictures with curious iconographies, such as the satyr form of Sagittarius (fig. 41), to an earlier book of Catasterisms on these grounds.⁵⁰ Whenever possible, the text of the De ordine and the imagery in the Carolingian Handbook of 809 emphasized meteorological details or made references to antiquarian forms of the constellations associated with zodiac cycles and celestial globes. These decisions were motivated by a zealous desire for scientific accuracy and an unshakable, fundamental belief in the intellectual sophistication of classical learning throughout the Frankish lands. The Franks could only achieve the scientific sacralization of the zodiac and the Greco-Roman history of the constellations by first fully immersing themselves in the historiographic tradition.

FIGURE 41

Lyra, Cygnus, Aquarius, Capricorn, Sagittarius, and Aquila. Staatsbibliothek zu Berlin, Stiftung Preußischer Kulturbesitz, Hs. Phillipps 1832, fol. 84v. Photo: bpk, Berlin / Staatsbibliothek zu Berlin / Art Resource, New York.

Christian Pictures of Pagan Constellations North of the Ecliptic

In Drogo’s copy of the Handbook of 809, forty-two descriptions⁵¹ of the constellations in single-column format are each followed by a free-standing Carolingian miniature. Drogo’s personal reading reenacted the passage of the celestial bodies, meditating upon the information about the stars presented in the trimmed-down, palatable Christian text, the De ordine. The paintings document the Carolingian reaction to painterly styles of modeled forms, which Frankish viewers could access in late antique mosaics, extant frescoes, and manuscripts such as the Calendar of 354, known to have existed in a Carolingian copy.⁵² The Carolingian star pictures, along with their linear style of execution, perhaps derive in part from such models, or from a lost late antique art of celestial globe production. Carolingian artists are known to have employed a polycyclic strategy of image selection when producing new pictorial cycles for codices that were intended to relay conceits and concepts linked to science and the liberal arts. The roots of the star pictures are nevertheless unequivocally linked to the history and textual transmission of the illustrated Aratean texts just discussed.⁵³ It is important to underline here exactly what is meant by an innovative or original pictorial cycle during the Carolingian period, especially vis-à-vis star pictures.

To make a general point, it is impossible to label anything even vaguely reminiscent of a crab a thoroughly original depiction of the zodiac sign Cancer. That is precisely because the form belongs to a standard type. The standardization of the form of the sign had been defined by human observation, in agreement with how people had long since clustered the stars in the sky into recognizable patterns. By the ninth century, the form of the constellation Cancer had been fixed for all of human time. The selection of the crab as an appropriate form for Cancer was no longer an interpretative artistic decision in the ninth century; its use provided an accurate scientific report of the facts of the matter. It would be just as meaningless to evaluate a chemistry student’s sculptural creativity with respect to her model of a water molecule, rendered accurately with three colored wooden balls and floppy springs. Mutatis mutandis, the same applies to contemporary digital imaging software, rendering discrete components of coded molecular structures more readily apparent and discernible than the blades of grass on the lawn beyond one’s window. In all these cases, the presentation conforms to an accurate model grounded in scientific observation and theory. Various star pictures become significant for the history of art when they are rendered with powerful individuality. The way in which the Carolingian artists in Metz represented the ninth-century images of the constellations in Drogo’s personal copy of the Handbook of 809 (Madrid 3307) turned the manuscript into a superior creative achievement. It is possible to evaluate Carolingian star pictures for their relative adherence to the prototypes derived from Aratean traditions and ancient star globes. It is equally possible to celebrate the Carolingian star pictures in Madrid 3307 for their innovative painterly forms of the constellations.

Beginning with the northern hemisphere, at the north pole, the serpent, Serpens, intertwines with the forms of the two bears: Ursa Major, otherwise known as Haelice or Arcturus Maior, and Ursa Minor, also known as Cinosura or Arcturus Minor. The forms of the miniatures in Drogo’s copy of the handbook derive in part from those found in the Aratean manuscripts, which were consulted in the composition of the De ordine in Corbie. The similarities and dissimilarities among the Carolingian star pictures become readily apparent when we compare the individual depictions of the bears and serpent (plate 1) with the single image of the trio found on folio 3v in the celebrated Leiden Aratea, which belongs to a set of manuscripts associated with the court of Louis the Pious (fig. 42).⁵⁴

The text begins with a description of the stars in the Great Bear: “Haelice, the Great Bear, has seven stars in its head. There is one star in each visible shoulder, one in the arm, one in the chest, two bright stars altogether in the front paw, one bright one in the tail, one bright one in the belly, two in the back leg, two in the rear paw, and three more in the tail. This makes a total of twenty-two stars.”⁵⁵ Immediately thereafter, an artist painted the Great Bear (plate 1). The descriptions of the constellations in Drogo’s book, derived in part from the descriptions in the Catasterisms attributed to Eratosthenes, had been stripped of their original mythological content.⁵⁶ In the text of the De ordine prepared for the Handbook of 809, the paramount concern was the composition of a text that would fulfill Charlemagne’s mandate for a pedagogical tool. Adalhard of Corbie was probably personally responsible for the removal of this mythological content from the Basel Scholia to rid the astronomical information of pagan associations.⁵⁷ This was perhaps the most important innovation made by the Carolingian author(s) of the De ordine, and it continued a practice that had already played an important role in the development of the Revised Aratus.

As Le Bourdellès has argued, the new Latin edition of the Aratean text from the late eighth century provided an ideal opportunity to render it up-to-date with the latest in Christian education—a goal that was in complete accord with widespread Carolingian strategies for scientific sacralization. Classical ideas in the Revised Aratus were intentionally juxtaposed with information culled from Isidore, and inspiration for at least one passage even harks back to the recently deceased Venerable Bede.⁵⁸ The suppression of the mythological material emphasized the condemnation of paganism by the prelates who had attended the synod of 809. The refusal to include or to examine the secular content traditionally associated with the constellations also underscored their rejection of late classical Roman paganism and the indigenous Germanic religious practices with which such beliefs intertwined in the ninth century.⁵⁹

FIGURE 42

Ursa Major, Ursa Minor, and Draco from the Leiden Aratea, Bibliotheek der Rijksuniversiteit, MS Voss. lat. Q. 79, fol. 3v. Photo courtesy of Universiteitsbibliotheek Leiden.

FIGURE 43

Arcturus Minor / Ursa Minor. Cologne, Erzbisch. Diözesan- und Dombibliothek, Cod. 83-II, fol. 155v. Photo courtesy of Erzbisch. Diözesan- und Dombibliothek.

Nevertheless, it is equally important to underscore that this Christian revisionist interpretative bent likewise permeated the artistic portrayals of the constellations in Drogo’s book.⁶⁰ For this reason, the requisite classical references are included in the discussion that follows, primarily to elucidate the areas in which Christian artists from Metz foisted an eighth- and ninth-century Carolingian outlook upon their star pictures and advanced a creative pictorial vision of the heavens in their novel reinterpretations of long-standing Aratean imagery.

Intriguingly, the miniatures in Drogo’s copy of the Handbook of 809 include neither representations of the adumbrated lists of stars nor a monumental format. The scale of the handbook’s miniatures is intimate, and the artistic refinement expressed by the painterly compositions was probably intended for the purposeful and pleasurable observation of its owner. Alternatively, Drogo’s copy may have been intended to be enjoyed in small group settings by a select number of elite students, either in the bishop’s company or in that of his academic masters at the cathedral school. It is easy to see how full-page versions of the same cycle of illustrations would have better fulfilled this purpose; however, it is equally meaningful to envision the use of Drogo’s manuscript in tandem with other astronomical observational tools appropriate to the ninth century, such as a star globe or a star clock (horologium nocturnum). Indeed, without such instruments, the lack of accurate stellar displacement in the images of the constellations calls into serious question their pedagogical efficacy.⁶¹ This is, however, to confuse the Handbook of 809 with a typical field guide for stargazers. The images are not arranged or presented in the same manner. The intellectual work for students of the liberal art of astronomy was different. Careful observation and analysis of the images of the constellations in Drogo’s copy of the Handbook of 809 on site in Madrid could not confirm Dieter Blume and his colleagues’ claim that the stars were once included in the constellation pictures of Drogo’s handbook.⁶²

The handbook fulfilled Charlemagne’s pedagogical desire for a treatise on astronomy and computistics in the service of reforms within the Frankish lands. The miniatures provided basic forms that could serve as mnemotechnic teaching aids for younger stargazers.⁶³ The lavish pictorial program presented in Drogo’s manuscript, however, surpasses in quality the linear drawings found in other copies of the Handbook of 809 (figs. 36–38). Even the simpler forms are adequately suited to fulfill their purpose as a teaching tool. The presence of such lavish miniatures in Drogo’s copy truly sets it apart as a royal commission worthy of an illegitimate son of Charlemagne. This heightened aesthetic quality gives Drogo’s book its apparent distinction as a definitive expression of the Handbook of 809. That does not mean that Drogo’s book is a more accurate reflection of any putative model; on the contrary, the profound difference in quality between Drogo’s book and the others underscores the unique nature of this creative undertaking by superior Frankish illuminators.

This lack of stars in Drogo’s handbook differs from the image cycle in the Leiden Aratea, which includes an original Germanicus Caesar translation. An additional Latin translation of the Aratea by Rufus Festus Avienus (fourth century) supplemented the text of the Germanicus translation in the Leiden manuscript.⁶⁴ In the image for the Great Bear, Little Bear, and Serpent on folio 3v of the Leiden Aratea, the three constellations are clustered into one picture without true concern for the text beyond the framed border (fig. 42).⁶⁵ This can be considered a general design strategy, which arranged visual information analogously to the semiotic construction of hypotactic subordinated sentences in Latin literature. This can be considered a form of visual hypotaxis, emphasizing integrated relational placement and expressing pictorial information with pristine clarity.⁶⁶ In a text displaying parataxis, such as a work by Sallust, ideas are linked together in clusters without the subordination of clauses.⁶⁷ In the descriptions and illustrations of the constellations from Drogo’s copy of the Handbook of 809, the De ordine presents a progression of text-image pairs. This sequence of syntactic text-image pairs remains subordinated to the overall structure of the De ordine, but follows the order of the constellations established by Aratus in a concatenated series of semiotic units, adduced in sequence like a paratactic history from Sallust. The imagery in Drogo’s book therefore separates the Bears from the Serpent in three discrete pictures (plate 1). The arrangement of the Bears and Serpent into three sequential text-image syntactic units emphasizes the form of each constellation and asserts the individuality of each one within the whole list of constellations in the De ordine.

This paratactic strategy was also adopted for the representations of the bears and serpent on folios 155r–156r of the Cologne Aratus (fig. 43). Comparisons of the types of images to be found in both manuscripts can tell us much about the kinds of lost late antique precursors that were harbingers of celestial understanding, relaying the classical traditions to Carolingian artists who then interpreted the ancient forms for themselves.

FIGURE 44

Hydra, Crater, and Corvus from the Leiden Aratea, Bibliotheek der Rijksuniversiteit, MS Voss. lat. Q. 79, fol. 76v. Photo courtesy of Universiteitsbibliotheek Leiden.

For example, three discrete images depict the Hydra, the Raven, and the Crater on folios 62r–v in Drogo’s copy of the Handbook of 809 (plates 16–17).⁶⁸ The Leiden Aratea represents the three clustered together in the framed miniature on folio 76v (fig. 44), while in the Cologne Aratus, the hydra, bird, and vessel are all included within one scene at the top of folio 167r (fig. 45). The artists working in Metz preferred to emphasize the individuality of the constellations and their significance for Carolingian readers by painting three miniatures. In Drogo’s book on folio 62r, one image clusters the trio together (plate 16), according to the more standard hypotactic form included in the Leiden manuscript, followed by individual depictions of the Raven and the Crater on folio 62v (plate 17). This conforms to the semiotic strategy of compilation through juxtaposition noted in the introduction. The painters of Drogo’s copy of the Handbook of 809 arguably sought to emphasize the coeval informational status of both text and image in the Handbook of 809. The painters in Metz pushed the design strategy of visual parataxis—experimented with in the Cologne copy of the Revised Aratus—even further. The upshot was that the pictures in Bishop Drogo’s manuscript were arguably more useful references of the forms of the individual constellations for teaching purposes than the miniatures available in the picture book, made for the court of Louis the Pious, contained in the Leiden Aratea.

The representations of the constellations of Serpens and the two bears in Drogo’s copy of the Handbook of 809 nevertheless recall the essential features of the corresponding animal forms in Leiden, taken one at a time. This was natural; the two cycles both derived from common pictorial prototypes, and of course depict the same constellations. The two bears look away from each other in opposite directions in both books. In the Leiden manuscript, they stand paw to paw across the divide delimited by the tortuous form of the serpent. In Drogo’s handbook (Madrid 3307), the bears both stand on all fours, heads forward in opposite directions. Whereas the bears in Leiden appear to leap through space in an endless, rhythmic circular chase, the two bears in Drogo’s book hover within the illusionistic space of the parchment folio. The bears in Madrid 3307 conform to the illusionistic pictorial logic appropriate to landscape painting, for they seem to stand on a missing horizon line. Medieval painters generally adopted the pictorial logic of the groundline but did not include one in most cases; in Madrid 3307, Heridanus on folio 61v may be an exception to this rule (plate 15).

More important, in Drogo’s handbook, as in the Cologne Aratus, the text block announces a spatial-semantic frontier that runs parallel to the implied horizon line on which the otherwise free-floating bears seem to stand in profile. In this way, the text not only provides the descriptions for the star catalogue but establishes a visual perimeter by blocking off the open space of the pictorial zone within which the ninth-century painters carried out their work. The fact that the paws of the Great Bear in Drogo’s book are set slightly above a line ruled in drypoint on the recto side of folio 54 provides additional visual proof that this layout was a design decision rather than an accident.

Unlike the Leiden Aratea images the miniatures in the Handbook of 809 do not have frames. The Carolingian artists who painted the original lost archetypal Handbook of 809 and the copy made for Drogo (Madrid 3307) dismantled the late antique framed images copied into the Leiden manuscript into the smaller discrete images preferred by the makers of the handbook. At the very least, the painters of Drogo’s book saw no reason to add any frames. Frankly, the preference for visual parataxis shattered the frames of the miniatures. This reduction in Madrid 3307 of the pictorial sequence into isolated units subordinated to the rhetorical structure of the De ordine was a preferred Carolingian pictorial book-making device, likewise confirmed by the Cologne copy of the Revised Aratus. The pictures in the Handbook of 809 deviate from the pictorial tradition copied by the artists who worked on the Leiden Aratea. As a result, the pictorial logic of the De ordine adheres to the twin Carolingian desires for (a) an original but simplified textual and pictorial presentation of the astronomical information appropriate to meditative study, and (b) perfect mastery in a pedagogical handbook.

FIGURE 45

Serpens/Hydra and Anticanis/Canis Minor. Cologne, Erzbisch. Diözesan- und Dombibliothek, Cod. 83-II, fol. 167r. Photo courtesy of Erzbisch. Diözesan- und Dombibliothek.

The lack of stars in the paintings of the constellations in Madrid 3307 is an odd solution, however, and calls into question the independent educational value of the miniatures. The simple text of the De ordine arguably permitted a general mental reconstruction of the placement of the stars within the painted image of the Great Bear. But such verbal descriptions could have benefited from the sort of visual representations of the stars within the constellations that are included in the Leiden Aratea. Alfred Stückelberger is the most optimistic voice about the reliability of the images in the Leiden Aratea, noting a strong correlation between the star catalogue of 1,022 fixed stars utilized by Ptolemy in the second century and the illustration of a constellation such as Sagittarius on folio 52v (fig. 21).⁶⁹ This probably overstated view has been significantly revised by Elly Dekker, who instead finds the general lack of accuracy in the Leiden Aratea star pictures to derive from the joint influence of both the original Ptolemaic star catalogue and diverse literary “descriptive” traditions, including the Basel Scholia under discussion here.⁷⁰

The upshot is, of course, that in both Drogo’s copy of the Handbook of 809 and the Leiden Aratea, there is only one potentially helpful source for the localization of the stars. There is an emphasis on the semantic potential of the text in Drogo’s book, which deviates from the preference for pictorially placed stars in the Leiden Aratea. There is no reason to believe, however, that the lack of stars in the depictions of the constellations originated with Drogo’s book. Presumably, the pictorial cycle designed for the original archetypal handbook created after the synod of 809 was made the same way. Interestingly, there are also no depictions of stars in the free-standing, single-column miniatures included within the early Cologne Aratus (fig. 8). Nor are there any frames around the illustrations on the two-column text pages in the copy of the Recensio interpolata (that is, the Revised Aratus explained above) manuscript from Saint-Gall, such as the ninth-century Saint-Gall 902 (fig. 40).⁷¹ As a result, the pictures in the original, archetypal Handbook of 809 probably resembled the cycles of pictures found in manuscript copies of the Revised Aratus, not the tradition of pictures associated with the Germanicus translation in Leiden. This only serves to demonstrate that the Handbook of 809 stood at the nexus of several programmatic efforts throughout the Frankish lands: the book simplified and comprehensively conveyed astronomical and computistical acumen to the schools; the pictorial program was a coeval presentation of period science and not merely a dilettantish fancy for prelates; and the Revised Aratus and the Handbook of 809 were twin efforts at developing a serious Christian presentation of classically inspired astronomical traditions and not merely the result of sanguine reclamation projects.

The representation of Ursa Major in Drogo’s copy of the Handbook of 809 follows its textual description on folio 54v (plate 1). The bear has a stern grimace and an upturned snout, as if it is on the defensive. Its paws terminate in black lines, which reduce the form to a stylized rendition of a quadruped, seemingly with hooves. This might simply be the result of the remaining underdrawing, however. A harsh black contour supplies the basic form of the animal. Certainly, some attention was paid to ursine anatomy, such as the convincing and elevated dorsal hump. Arguably, there is a sense of overlapping perspective in the representation of the limbs of the Great Bear. In fact, the underdrawing has bled through the pigment in the right foreleg and hind leg, suggesting that the artist employed underdrawing to sketch in the animal’s general form. The painter nevertheless manipulated the basic form of the animal at a second stage, while applying the color to the parchment. This is highly informative about the practices of Carolingian painters working in Metz. Each artist first created an outline drawing before completing a miniature. Since the styles of painting and the styles of the underdrawing always display a high level of coordination, it is highly likely that each painter also provided his own preliminary contour drawing. The sketch expressed the basic form, while the finished figure communicated the mature aesthetic vision of the artist. Such a two-step process argues in favor of a sophisticated Carolingian painterly practice that had already developed by the 830s, when Drogo’s copy of the Handbook of 809 was created.

The almost mirror image of Ursa Major, Ursa Minor (plate 1)—also known as Cinosura, according to the Epitome Catasterismorum that is derived from the now-lost star catalogue with myths contained within the original Catasterisms of Eratosthenes,⁷² and also as Arcturus Minor—reproduces the overall form of its bigger relative in reverse: “Cinosura, Ursa Minor, has four bright stars on one side arranged in a square shape; there are three bright ones in the tail. They total seven. Under these the star appears that is called Polaris, around which it is believed that the whole sphere is turning.”⁷³ In keeping with the Carolingian reformers’ aim to cleanse the descriptions of the constellations of pagan mythological associations, the only reference to the role the bears played in saving Zeus from Cronus lingers in their alternate names. According to the legend, Haelice and Cinosura attended to the infant god for a year in a cave near Mount Ida at Lyctus. In addition, the Carolingian author omitted any reference to the shipping practices of the Greeks, who navigated with the assistance of Ursa Major, and removed the reference in the Phaenomena to the superior navigational skills of the Phoenicians, who relied instead on Ursa Minor. The little bear reportedly revolves in a tighter orbit around Polaris, yielding a lower relative error for Sidonian sailors.⁷⁴

The loss of the navigational anecdote from the Phaenomena is a significant one. Navigation was one of the most celebrated uses of the constellations from antiquity into the medieval period, and there was no obvious reason for the clerics to rid the star catalogues of these pragmatic applications. This removal was a deliberate hermeneutic maneuver. The Handbook of 809 demanded a list of star pictures. Telling time was a practical concern of Frankish ecclesiasts such as Drogo, whereas the benefits of the constellations to Phoenician seafaring were not a priority within the cloister walk or cathedral precinct.

Like Ursa Major, Ursa Minor is brown with black highlights. Especially noteworthy are the efforts at rendering the paws, and the eye delineated by a heavy lid under a harsh eyebrow and a beady pupil. Although there was an effort made to capture the ursine form in both images of the bears, neither displays an overall attention to the texture of their coats.

The next constellation in the series is Serpens (plate 1), which weaves between the bears near the north celestial pole. This image of the constellation Draco resembles the style of other serpents painted in Metz during the ninth century, as noted in chapter 1. Since the archetypal Handbook of 809 arose in the court at Aachen in the years following the synod, 809–12, the style of the archetypal paintings would undoubtedly have reflected the stylistic trends associated with a later phase of the development of Charlemagne’s court school. The juxtaposition of three blue tonal values suggests the rounded form of the serpentine flesh. First, a light blue wash laid in the base color of the image, and then the Carolingian painter adapted a technique common to late antique fresco painting to model the forms.⁷⁵ The progressively darker shades of blue indicate shadows on the serpent, but also approximate the scaly texture of the reptile: “The serpent which lies in the space between the bears has five bright stars in its head and ten stars throughout the whole body. They total fifteen.”⁷⁶

It is important to underscore that although both the text of the De ordine and its imagery in Drogo’s copy of the Handbook of 809 hark back to earlier Aratean traditions, the text itself could not provide a painter with the pictorial cycle. For the pictures in the Handbook of 809, the artists had to look to illustrated Aratean manuscripts or other illuminated codices available in the ninth century. Since star pictures regularly supplied visual summaries of the information included in their texts, odd iconographic details could be carried over between manuscripts when old pictures were consulted in illustrating new texts. Sometimes artists employed stylistic devices (such as the stippled modeling of the serpent’s flesh) that could not have been derived from the text. Texts exist in visual cultures, too, and artists executing their commissions naturally draw on a multiplicity of sources. For this study, I emphasize Carolingian astronomical manuscript illumination for the sake of economy and to clarify meaningful aspects of the Frankish contribution to early medieval astronomical illustration.

The painters in Metz used the scoring on the opposite side of the handbook’s folio 54v as a guide for the undulating twists of the serpent’s body. This was a common practice in the atelier making these star pictures, and a similar layout recurs throughout the pictorial program. The form of the serpent adheres to a standard twisting shape also seen in the Leiden Aratea, despite the tendency of painters in Metz to employ their style of visual parataxis, eliminating the bears. The space remaining at the bottom of folio 54v, however, permitted the introduction of an additional bend in the snake’s body. This reveals the extent to which the pictures in Drogo’s handbook were made by creative painters who made ad hoc, innovative decisions rather than merely introducing all of their images copied in rote fashion from diverse sources.

The next image, on folio 55r, depicts Hercules (plate 2). The painter modeled the figure of Hercules with three ochre flesh tones. The use of three shades resembles the technique for modeling the scaly flesh of the serpent on folio 54v. The hero’s protruding, heavy brow and pupils resemble those of Ursa Minor. These miniatures can be attributed, then, to a first hand. The text explains that “Hercules, of whom it is said that he is kneeling, has one star in his head, one in his arm, one bright star in each shoulder, one star in the left elbow, one in the same hand, one in each hip, two in the right leg, one in the foot, above there is one in the club in the right hand, there are four stars in the lion pelt. They total sixteen.”⁷⁷ The image of Hercules includes a visual reference to the Aratean character known as the kneeling one, or “Engonasin,” here dubbed “Ingeniculo.” Aratus explained that the figure tilled the earth, like a farmer.⁷⁸

As Florentine Mütherich has noted in her discussion of the miniatures in the Leiden Aratea, the connection between Hercules and the kneeling figure was made at an early time and pervades the imagery of the Aratean manuscripts. A Revised Aratus manuscript, such as the Cologne Aratus, instead represents Hercules attacking the Serpent or Hydra (Draco), who is wrapped about the Tree of the Hesperides (fig. 8). This scene better illustrates certain scholia attached to the Germanicus translation of the Phaenomena.⁷⁹ The artists who created the pictorial cycle for the Handbook of 809 removed the mythological battle, while keeping the older visual reference to the “one who kneels.” In Drogo’s copy of the Handbook of 809, the painter decided to portray Hercules from the front rather than in the quasidorsal view displayed by the miniature in Cologne.

The visual references to the defeated Nemean lion and the club of Hercules gave the painters in Metz captivating narrative details to portray. In this respect, the image seems at odds with the desire to create a sterilized, Christian view of Hercules. Be that as it may, there remained a need, in compliance with the pedagogical goals of the original handbook, for the artists to render a form recognizable as the standard type of Hercules/Engonasin. The educational aims of the handbook vied with the clerics’ desires to sanctify the pictures, precisely because the classical content of a figure such as Hercules could not be separated from its form. The description of Corona precedes an image (plate 2) of a laurel with a jeweled clasp: “The crown has eight stars placed in a circle of which three are bright ones that abut the head of the northern serpent.”⁸⁰ Curiously, the artists painted the laurel in the same ochre hues as their depiction of Hercules. Green was used later—for the image of Piscis Magnus on folio 61v (plate 15), for example—so this hue was available to the painters in Metz. This suggests that these miniatures with the brownish hues at the beginning of the handbook are all the work of the first painter. The economical reuse of the pigments for the laurel crown and for Hercules (who is stylistically related to the image of Ursa Minor) reinforces the idea that all the paintings thus far are by a single hand. Although stylistic analyses of scientific manuscripts are not a common concern for historians of either art or science, this has led to the inaccurate dismissal of the qualitative richness of certain scientific or philosophical manuscripts, and to ignorance about the sorts of teams that were manufacturing medieval books that were not overtly biblical for meditation and study.

The next image also conforms to the design strategy of visual parataxis (plate 3).⁸¹ It is true that the Serpentarius, “the man wrapped about with a serpent,” also known as Ophiuchus, would be meaningless without his reptilian adversary in the picture:

Serpentarius, who is called Ophiuchus by the Greeks, has one bright star in the head, one bright star on each shoulder, one bright star on each foot, three stars in the left hand, four bright stars in the right hand, a single star in each elbow, a single star in each knee, a bright star in the right leg. They total seventeen.

The serpent, which he holds in his hands, has two stars in its mouth and four dim ones in the head, although it has two stars in the area up to the hand, and it has fifteen stars in the bend of the body. They total twenty-three.⁸²

But here, too, the artists deviate from a type of Aratean image that portrays Serpentarius trampling Scorpio, as on Leiden folio 10v, for example. Serpentarius and Scorpius (plate 3) each have their own miniatures in Drogo’s copy of the Handbook of 809, according to the pervasive paratactic design preference for a single picture per constellation. The artists who had made the cycle of images for the computistical manuscript in Cologne also separated out the pictures, accordingly, into two scenes on folio 157r–v of the Cologne Aratus:

Scorpius has two stars on each claw, or preferably both lips, before which there are three bright ones, of which the one in the middle is brighter; there are three bright stars on the back, two in the stomach, five in the tail, and two stars in the spine. They total nineteen. Before these stars four more were placed, whether they are considered to be in the claws or the lips, which are called the “chele of Scorpio,” and that were assigned under the circumstances to Libra. On account of its size it [Scorpio] was in two houses of the zodiac. So, it [Scorpio] was divided into the space appropriate for two zodiac signs.⁸³

The history of the zodiac is divided into two discrete phases. In the early phase, around the late sixth century B.C.E., there were eleven accepted signs of the zodiac. Scorpio and Libra were conflated into a large-clawed creature reminiscent of a lobster. This tradition harks back to such ancient astronomers as Anaximander (d. 546/545 B.C.E.), Cleostratus of Tenedos (active ca. 500 B.C.E.), and Oenopides of Chios (active in the second half of the fifth century B.C.E.). After Alexander the Great (d. 323 B.C.E.), the twelve-part zodiac wheel gradually became normative throughout the Mediterranean.⁸⁴

The introduction of a pithy tidbit from astrological history into the otherwise terse descriptions of the constellations is intriguing. The scientific writers from Corbie apparently found the fact noteworthy. Two points should be noted here. First, the Carolingian author was aware of not just the standard twelve-part zodiac but also of the alternate ancient eleven-part zodiac system. This suggests that the ninth-century writer of the De ordine had a sophisticated understanding of the history of astrology. Second, the Handbook of 809 as a whole was first and foremost an astronomical-computistical treatise. The author of the De ordine emphasizes the relationship between the twelve months of the year and the corresponding twelve-part series of zodiac signs associated with the calendar, and is also aware to some degree of the problematic impact of equinoctial precession on such alignments.⁸⁵ Depictions of the months and their related signs in Western medieval manuscripts date back at least as far as the Calendar of 354, and calendrical illustrations, such as a cycle of the months (consider fig. 22), embellish manuscripts to the end of the late Gothic period.⁸⁶

The image of Scorpius/Scorpio⁸⁷ in the handbook includes the long pincers mentioned in the text, and it therefore alludes to both the standard form of Scorpio and arguably to its more ancient prototype (belonging to the eleven-part zodiac cycle). This was also the form of Scorpio adopted for inclusion in the Revised Aratus from Cologne on folio 157v. In Drogo’s copy of the Handbook of 809, the dark brown and black underdrawing yields a sophisticated reticulation of the arachnid’s form.

The next painting in Drogo’s handbook, on folio 56r, is of Boötes, otherwise known as Arctophylax (plate 4). According to Aratus, Boötes endlessly prods the circumpolar constellations of the Great and Little/Lesser Bears as they circle the north pole, which is why the Carolingian painter presents him with his shepherd’s crook: “Boötes, who is called Arctophylax by the Greeks, has four stars in his right hand that do not set, and in his head he has one bright star. He has one star in each shoulder, one star per pectoral muscle with the brighter one on the right, beneath one pectoral muscle there is a light star. In the right elbow there is one bright star. Between both knees there is a bright star, and there is one star per foot. They total fourteen.”⁸⁸ There are losses to the head of the figure in Drogo’s book that probably took place at the time of the seventeenth-century rebinding (discussed at the end of chapter 1), when the early modern bookbinders cropped the folios. Enough of the brownish flesh-tones and blue mantle remain to warrant an attribution to the first painter. The characteristic large eye resembles other eyes found in the miniatures of Hercules and Ursa Minor.

This first artist also painted the angelic vision of Virgo (plate 4): “Virgo has one light star in her head, one star per shoulder, on the left upper arm/wing she has one light star, on the right upper arm/wing she has one star, she has one star per elbow, one star per hand, and on the left there is a brighter star called ‘the Spica, the ear of grain,’ in the tunic there are six dim stars, there is also one star per foot. They total eighteen.”⁸⁹ The maiden Dike is referred to by Aratus in the Phaenomena and is alluded to in the Carolingian description of Virgo.⁹⁰ The winged goddess of Justice, who retreated from earthly life according to the mythological tale, was arguably reinterpreted in an angelic form by the Christian painter. As the winged harbinger of justice, the mythological presentation of Virgo held her ear of grain or sheaf in her left hand, according to the text. This is highly significant, because Carolingian representations of the constellation Virgo differed greatly. She holds three sheaves of grain in her right hand in the Basel Scholia on folio 18v, whereas she holds scales in her role as a personification of Justice in the manuscripts of the Recensio interpolata such as Saint-Gall 902, page 86.⁹¹

The conscious removal of the scales, which identified Virgo in her mythological role as winged Justice, permitted the painter of the maiden in Drogo’s handbook to sanctify the representation of the zodiac sign. This innovative form removes the pagan mythological content from the image of Virgo, just as the author of the De ordine in Corbie made every effort to omit mythological references from the text. This unified editorial vision provides the best evidence for the hypothesis that Adalhard of Corbie was instrumental in the development not only of the text but also of the pictorial program for the original Handbook of 809.⁹² The lack of scales is also an intelligent scientific reaction to the content of the previous passage treating Scorpio: since the scales qua Libra had been removed from Scorpio, it would have been odd to pass them along to Virgo and include them in either her description or her depiction.

The salmon hue used for the overtunic worn by Virgo is also used for the mantles worn by the curiously sexless (censored) Gemini twins (plates 4–5). Granted, gender-neutral figures are common in early medieval art. The relative uniqueness of this decision in Drogo’s copy of the Handbook of 809 becomes apparent, however, when that depiction of the nude Gemini is compared with other copies of the Handbook of 809 displaying sexual organs (figs. 36, 38). The manner of painting the eye sockets makes it clear that this image was also painted by the first hand. The faces of the Gemini twins, especially the one on the left, recall the linear construction of the face also found in the image of Cepheus on folio 57v (plate 7). This early medieval technique for rendering the face of the figure requires a z-shaped line that begins with a heavy brow, descends as it defines the contour of the nose, and then ends with a line that defines the bottom of the nose. Two tick marks frame the philtrum of the upper lip. One or possibly two roughly parallel lines define the lower lip and seemingly dimpled chin of figures like these in early medieval manuscripts. When a figure stands in a three-quarter pose facing left, like the left-hand twin of the Gemini, the z-shaped line begins on the proper right side of the figure’s face, corresponding to the viewer’s left (plate 5). Examples of figures looking toward the right with such conventionally rendered faces include Virgo on folio 56r (plate 4) and the right-hand Gemini twin on folio 56v (plate 5):

One of the Gemini twins, who is next to Cancer the Crab, has one bright star in his head, and one bright star per shoulder. In the right elbow he has one star. In each of the knees there is one star, and in each of the feet there is one star. They total eight.

The other one has one star in his head, one star per pectoral muscle, and two stars in the left elbow. He has one star at the tip of his hand, one star in the left knee, one star per foot, and one beside the left foot that is called Propus. They total ten.⁹³

The image of a standing soldier and musician was atypical for Carolingian depictions of the Gemini, although not a unicum, and the symbolic reference to the depicted set of twins can be explained by appeal to the Leiden Aratea. The related image in the Leiden Aratea on folio 16v (fig. 46) nevertheless switches the left-right orientation of the twins in the miniature from Madrid 3307 (plate 5). The Gemini twins also look away from each other in Drogo’s copy of the Handbook of 809, whereas they engage each other within the visual confines of the square frame in the Leiden Aratea. As for the image of Hercules, the classical content of the legend provides the iconographic source for the depiction of the twin children of Zeus who erected a strategic wall for Thebes.⁹⁴

Whereas in the Leiden image one of the twins, Zethus, holds a lance resembling a spear and rests upon his club, he only holds the lance in the miniature from Drogo’s copy of the Handbook of 809.⁹⁵ The image in Drogo’s copy draws in a conservative way upon the mythological information that provided the distinguishing emblems of the Gemini. In this case, the imagery of the lyre denotes Amphion as the musician, and armed Zethus is the warrior.⁹⁶ The text of the Cologne Aratus refers back to an alternate and common set of names for the mythical brothers, Castor and Pollux, on folio 158r.⁹⁷ This text makes the more common identification of the Gemini twins as the “Dioskouroi.” The image in the Leiden Aratea also includes an allusion to Castor and Pollux. The brothers’ white domical caps are emblems of the Dioskouroi that the painter sacralized in the Leiden Aratea picture, setting gold Christian crosses on top of their headgear (fig. 46).⁹⁸ This editorial act is symptomatic of the general approach to reform with respect to the images and ideas linked to the liberal art of astronomy. Occasionally, a Christian gloss was added to a classical concept, but equally importantly, the classical content in the original needed to be elevated to a careful, reflective level of significance before it could be revised. Exegetical emendation already embeds a tactical respect for, or fear of, the original into the very idea of revision.

Each time a new manuscript was created, these images, and their lost precursors, were assessed for their potential inclusion in the books associated with the various Aratean recensions. The situation became even more complicated when monks combined or altered certain cycles of star pictures, seemingly ad hoc, recombining favorite exempla, or pulling together scant resources, in order to create an astronomical or computistical handbook or encyclopedia anew. It is vital to distinguish between moments of creative interpretation and the occasional effort at straightforward classical emulation.⁹⁹

The first painter in Drogo’s copy of the Handbook of 809 also rendered Cancer the Crab with a convincing anatomical form (plate 5). Tonal washes complement a confident contour line and model its features:

Cancer the Crab has two bright stars in its chest that are called the Donkeys between which there is a little haze, which seems to be of a bright color and that they call the manger. Cancer has one star per foot on the right side, all four are dim. On the left side, Cancer also has two bright stars in the first foot, two in the second, one in the third, and a little one in the fourth foot. In the right claw or lip Cancer has three stars, but it has two stars on the left. They total seventeen.

From among this group, the two stars, which we set in the chest of Cancer with two small and dim stars, are called the Donkeys. So, they total four stars.¹⁰⁰

The references to the two donkeys in the text and the miniature seem odd at first. In fact, by including the donkeys in the image of Cancer, the author of the De ordine further demonstrated his integral involvement with the creation of all of the texts compiled into book V. One northern and one southern star designated as the donkeys gather at the manger, which lies before Cancer, according to Aratus in the Phaenomena. Both stars were considered useful for predicting inclement weather. For example, a fictitious rhyme can easily pass for an old adage: “if the manger grows dark while the stars twinkle, it surely soon will sprinkle.” In fact, according to Aratus, the donkeys were useful for making precisely such unimpressive meteorological predictions, which nevertheless would arguably benefit travelers and agriculturalists: “Observe also the Manger . . . if suddenly, when the sky is clear all over, it disappears completely, and the stars that go on either side appear close to each other, then the fields are drenched with no mean storm. If the Manger should darken and the two stars be at the same time recognizable, they will be giving a sign of rain.”¹⁰¹

FIGURE 46

Gemini. Leiden Aratea, Bibliotheek der Rijksuniversiteit, MS Voss. lat. Q. 79, fol. 16v. Photo courtesy of Universiteitsbibliotheek Leiden.

Alternatively, when the northern star darkens but the southerly one remains fully visible, the winds rise in the south, and vice versa for the southern star.¹⁰² The information to be gleaned from this is that when winds are blowing clouds and fog in a given direction, the observer notices things getting blurry. This is not particularly helpful, but it does offer an empirically oriented natural explanation for precipitation and changing visibility that does not require a pantheistic view of the environment. To my knowledge, none of the earlier Aratean manuscript traditions that include pictorial cycles show the donkeys feeding at their manger before Cancer the Crab.¹⁰³ The image of Cancer in the Handbook of 809 relies on the ancient Aratean text in the Phaenomena, and on Pliny’s recapitulation of the same information in book V.12 of the Handbook of 809, as discussed further in chapter 4. The creator of this innovative image thereby also demonstrated his astute awareness of late antique and early medieval science. This reinforces the suspicion that Adalhard was involved in the selection of images for inclusion in the original handbook. In fact, Le Bourdellès follows Bernhard Bischoff and has determined that a dismembered copy of Pliny was an early ninth-century copy from Corbie. As Le Bourdellès argues, Adalhard would have been familiar with this material.¹⁰⁴ This further buttresses the likelihood that Adalhard played an important role in the iconographical innovation for the images in the handbook.

Early Christian astronomy had considered certain zodiac signs of meteorological importance for the agrarian and seafaring cultures of the Mediterranean. The significant texts about meteorology (taken in the broadest sense to mean any scientific assessment of the effects of weather) through 400 c.E. included Pliny’s Naturalis historia, already noted; Columella’s De re rustica (first century);¹⁰⁵ and the translated Latin versions of the Aratean texts, the Phaenomena and the Prognostica.¹⁰⁶ Interest in the influences that celestial bodies could exert on meteorological matters waned during the sixth and seventh centuries. Stargazing continued to be useful, however, for computistical calculations, which were necessary for the creation of the church calendar, as addressed in chapter 1. Two texts in particular were instrumental in perpetuating this revised vision of the utility of astronomy for Christian communities: the Institutiones divinarum et saecularium litterarum (Institutions of divine and secular learning; hereafter Institutiones), by Cassiodorus (d. 585) in the second half of the sixth century, and the De natura rerum of Isidore of Seville in the early seventh century.¹⁰⁷ (These texts are also discussed in chapter 4.) In this light, the allusions to meteorological concerns in the Carolingian De ordine were actually a creative revival of early Christian interests.

The first painter likewise rendered the image of Leo using his preferred ochre palette (plate 6): “Leo has three stars in its head, two in its neck, one in its chest, three along the back, one in the middle of the tail, one bright star at the tip of the tail, two under the chest, one bright star in the front paws, one bright star under the stomach, one bright star in the middle of the belly, one in the groin, one in the back knee, one bright star in the first paw. They total eighteen. Seven other stars next to its tail seem relatively dim.”¹⁰⁸ Nothing mattered to the Carolingian painter here besides rendering a convincing portrayal of the constellation associated with the zodiac sign Leo. All references to the Nemean lion slain by Hercules and to the animal’s role as the legendary king of the beasts have been suppressed.¹⁰⁹

In the next miniature on folio 57r, however, the Carolingian painter included depictions of the Hedi, or goats. They are two stars said to be near the left hand of Auriga, the charioteer (plate 6): “Auriga, otherwise known as the charioteer Ericthonius, has one star in his head and one star per shoulder; the brighter named, the she-goat, is on the left. Each knee has a star. At the tip of the [right] hand are two stars and in the left hand two more that are called the ‘Hedi,’ the Goats. They total nine.”¹¹⁰ The image from the Leiden Aratea also displays the goats in the left hand, but without the team of two horses, on folio 22v. The blue and brown steeds in Drogo’s copy of the Handbook of 809 are fanciful, but they display a generally convincing equine form. The drawing of the goat—which is partly in the margin, rather than over Auriga’s left shoulder—actually resembles the donkeys accompanying the image of Cancer. The donkeys of Cancer, the goats of Auriga, and the horses pulling Auriga’s chariot all share a robust abdomen and attenuated limbs, which terminate in similarly stylized hooves. These forms were all made by the handbook’s first painter. The tunic and mantle of Auriga come from his limited palette: blue, salmon pink, ochre, and brown.

Two interesting details emphasize the desire of the Carolingian author and painter to accentuate the scientific content of the description of Auriga. First, according to Germanicus, the character Ericthonius, referred to in the description, allegedly came up with the idea for the four-horse chariot. Germanicus alternatively suggests that the constellation could refer to the mythical assassin of King Oenomaus, Myrtilus.¹¹¹ By focusing on Erichthonius, the Carolingian author and artist of the De ordine drew attention to technological advancements in equine transportation rather than infamous examples of regicide. Secondly, the stars known as the goats, like the donkeys accompanying Cancer, were included in the description because they were considered useful for predicting inclement weather. In the Phaenomena, Aratus explained that the Olenian goat, which had nursed Zeus, rested on the shoulder of Auriga, while her kids were the two goats on the charioteer’s left hand. These two stars appeared dim whenever the sea raged.¹¹² The wet nurse of the king of the Olympian gods, the Olenian she-goat, is relegated to the margin of Drogo’s ninth-century manuscript, while her kids are represented in the center of the leaf without reservation. The reason was twofold: symbolically, the she-goat included a suspect mythological reference, whereas the kids were useful on pedagogical grounds. Practically speaking, however, the composition called for the inclusion of the kids in the center of the miniature. The Olenian she-goat was only useful for the composition insofar as she drew attention to her kids.

The same painter supplied the bucolic and rather lifelike representation of Taurus on folio 57v (plate 7). The eye of the bull resembles the eyes of the other animals in Drogo’s copy of the Handbook of 809. An interesting departure in this image, however, is the use of the salmon hue as a fleshy highlight. In the original cycle of the handbook, the full-length reclining image of Taurus deviates from the standard halffigure found in the Leiden Aratea on folio 24v and in the Recensio interpolata manuscript, Saint-Gall 902, on page 87¹¹³: “Taurus has one star in each horn, one star in each eye, one star in the nose; these five are called the Hyades. In the hoof Taurus has three stars, two in the neck, two in the back, one final brighter star below the stomach, and a bright star in the chest. They total fifteen. There are seven of the stars, which are called Atlantides or Pleiades; of these six can be seen, but the seventh is faint. It is said that they were placed in the tail of the bull, Taurus.”¹¹⁴ The images painted in Drogo’s copy of the Handbook of 809 and perhaps its archetype were more likely than not either an original appropriation from zodiac cycles of the complete bovine form or a truly innovative attempt at rendering a bull. The reference to the Hyades is another example of a reference to meteorological lore. In the classical world, the ascension of the Hyades in conjunction with the sun was considered a sign of rain.¹¹⁵ The inclusion of the Pleiades under the description of Taurus demonstrates another authorial effort to omit unnecessary mythological references. While the Leiden Aratea includes an image of the seven daughters of Atlas on folio 42v, the originator of the pictorial cycle in the Handbook of 809 excludes any portrait of the mythological siblings. In the Christian Carolingian Handbook of 809 meteorological prognostication with the stars was useful, but mythical stories about the daughters of Atlas, the Pleiades, were not. Consequently, the author of the De ordine found a way to include the reference to the seven stars that usher in fall or spring while downplaying their mythological content.¹¹⁶

FIGURE 47

Ascension. Drogo Sacramentary, Paris, Bibliothèque nationale de France, MS lat. 9428, fol. 71v. Photo: BnF.

That these explanations are not themselves included in Drogo’s book is also significant. The labels, names, and pictures of constellations or star clusters in the Handbook of 809 were mnemotechnic, insofar as they enabled a Carolingian student to create a taxonomic, hierarchical inventory of astronomical information and folk tradition. The cursory descriptive quality of the text of the De ordine parallels the uninformative nature of many of the star pictures when they are considered in isolation. The Handbook of 809 offers introductions for pupils and discussion points for teachers, yielding an elementary synthesis with royal imprimatur.

The next three images in Drogo’s copy of the Handbook of 809 were also painted by the first artist working on that manuscript (Madrid 3307). They depict the characters immortalized by the mythological tale of Andromeda’s rescue from the sea monster by Perseus (plates 7–8). None of the descriptions makes this clear:

Cepheus has two bright stars in his head, one bright star in his right hand, one dim star in his elbow, one bright star in the left hand, one in the left shoulder, three stars arrayed obliquely in the baldric, two stars on the right hip, two stars on the left knee, one star at the tip of the foot, three stars above the foot. They total seventeen.

Cassiopeia has one bright star in the head, one bright star per shoulder, one bright star on her right breast, one bright star on her right thigh, two bright stars on her left thigh, one bright star in the knee, and there are two bright stars in the pedestal of the chair where she sits. They total ten.¹¹⁷

Andromeda has one brilliant star in her head, one star per shoulder, one star in the right elbow, one bright star in the hand, one bright star in the left elbow, one star in the forearm, three stars in the girdle, four more above that area, one bright star per knee, two stars in the right foot, one star in the left foot. They total nineteen.¹¹⁸

Each figure displays the same pronounced eye sockets and limited range of colors, making the most of blue, salmon, and yellow-brown. These three miniatures betray, however, a quality of color that is thin and diffuse, like washes. The quality of the painting almost tapers off at this point, as though the pace of production had accelerated. Yet this use of painterly washes has often been associated with the style of painting in Metz during the ninth century (figs. 33, 47–48).

FIGURE 48

Crucifixion. Drogo Sacramentary, Paris, Bibliothèque nationale de France, MS lat. 9428, fol. 43v. Photo: BnF.

This first painter of Drogo’s copy of the Handbook of 809 was adept at rendering animal forms such as that of the fabulous winged horse Pegasus on folio 58r (plate 8). Its eyes resemble those of Taurus (plate 7). The hooves and equine grimace also resemble the second, brown horse in the team pulling the chariot of Auriga (plate 6): “The horse called Pegasus has two bright stars in its face, one star in its head, one in the haunch, a bright star in each ear, four stars along the neck, the next bright star is the [aforementioned] one in the head, has one star in the shoulder, one in the chest, one bright star in the navel, one star in each of two knees, and stars in each of those hooves. They total seventeen.”¹¹⁹ The truncated representation of Pegasus resembles the image found in other manuscripts derived from Aratean precursors.¹²⁰ It is interesting that there is a tight relationship between the description and depiction of Pegasus. The text clearly relates to the frontal presentation of a bifurcated animal, since a complete equine form would have called for at least four more stars, for a total of at least twenty-one. In contrast, the Pleiades set within the tail of Taurus necessitated the inclusion of an entire recumbent bovine form. Arguably, the choice of animal picture here derived at least in part from an assessment of the total number of stars and ideas expressed by the star catalogue, the De ordine, in Drogo’s copy of the Handbook of 809. The fact that Pegasus is the only half-length figure in Drogo’s Handbook of 809 confirms this and underscores the book’s educational importance for the Carolingian study of astronomy.

The first painter employed the same painterly technique for rendering the hooves of the ram Aries (plate 9) that he used for the hooves of the Olenian she-goat in the Auriga miniature (plate 6). The blue wash coloring the ram’s fleece is girded by a green laurel. This symbolizes the ram’s role as the inaugurator of a new zodiac cycle and as the zodiac sign initiating spring.¹²¹ In other words, the imagery once again takes full advantage of an extant pictorial tradition to emphasize those aspects of the Handbook of 809 that clarify the relationship between the constellations and other celestial phenomena, such as seasonal shift, permitting the straightforward Christian association of springtime renewal with Easter. At this point in the text, many of the descriptions become terser: “Aries has one star in the head, three stars in the nostrils, two in the neck, one bright star at the tip of the right foot, four in the back, one in the tail, and one star at the tip of the left foot. They total thirteen stars.”¹²² The salmon-colored triangle Deltoton follows in sequence (plate 9): “The triangle, which is called Deltoton by the Greeks, has three stars altogether with one in each vertice.”¹²³ The image of Pisces displays a celestial thread of stars linking the mouths of the pair of fish known as Borius/Aquilonalis and Australis (plate 9): “There are two Pisces fish, of which the northern one (Borius) is called Aquilonalis and has twelve stars; the southern one (Notius/Notus), however, is called Australis and has fifteen stars. They are joined together with a line, which has three stars toward the northern part, three to the south, and three stars toward the west. The total number of stars is thirty-six.”¹²⁴ Pisces is the final zodiac sign through which the sun passes on its annual journey and designates the conclusion of the cyclical progression of the seasons. In some Aratean images linked to the Germanicus translation, the fish are joined tail-to-tail with a knotted set of chains that restrict their mobility and reportedly cause them to fight for dominance as they swim in opposite directions, such as in the image from the Leiden Aratea on folio 38v.¹²⁵ The image in Saint-Gall 902 belonging to the Recensio interpolata displays instead the fish joined by a cord at the lips on page 90, as they are likewise depicted in Drogo’s copy of the Handbook of 809. In what remains of the damaged miniature in Drogo’s copy, the three-tone technique has been used by the first painter to model the form of the fish. The feathery tail fin lacks a certain anatomical verisimilitude, but salmon-colored highlights added pectoral fins and relatively convincing fleshy sets of gills. Apparently, there was no normative guideline that regulated the placement of the cord linking the fish.¹²⁶

Perhaps the first painter relied on an image of Pisces from a zodiac wheel when painting the representation in Drogo’s copy of the Handbook of 809. For example, the fish are joined at the mouth in the zodiac from the Utrecht Psalter, folio 36r (Utrecht, University Library, MS 32, fol. 36r; fig. 49), although in that depiction Aries was drawn without its wreath. The belt, wreath, or hoop around the ram’s waist (plate 9) indicates the position of the constellation at the equinoctial colure, or the astronomical dividing line, which crosses the ecliptic at the equinoxes and separates the fall/winter from the spring/summer seasons. Aries is the first sign of spring at the vernal equinox, and by jumping through the circular celestial ring, or “hoop,” uniting the equinoxes, the ram ushered in the spring and summer signs of the zodiac.¹²⁷ This underlines the extent to which the basic forms of the zodiac were established, but iconographic details occasionally supplied additional nuances of meaning. The representation of Pisces in Drogo’s copy of the Handbook of 809 exemplifies the conventional form associated with zodiac wheels. Arguably, this conscious decision underscores the placement of Pisces at the end of the season of death and dying. The image of Aries with its hoop referred to the renewal awaiting nature in the spring and summer months of the year. Symbolically, as in the image from the Utrecht Psalter, Aries ushered in the time of redemption and spiritual renewal, which is one reason Christ triumphs over the lion and serpent at 12:00 in the zodiac from the Utrecht Psalter (fig. 49). These iconographic subtleties would have been appropriate for inclusion within a Christian computistical compilation such as the Handbook of 809, emphasizing the links between the zodiac, their constellations, and Easter.

FIGURE 49

Zodiac for Psalm 64. Utrecht Psalter, Universiteitsbibliotheek Utrecht, MS 32, fol. 36r. Photo: Universiteitsbibliotheek Utrecht.

An assistant to the first master painter finished the three images on folio 59r of Perseus, Lyra, and Cygnus (plate 10). The profile image of Perseus resembles Serpentarius on folio 55v (plate 3), seen from the side and back. The musculature of Perseus recalls that of Hercules on folio 55r overall (plate 2), with ochre shadows modeling the human form. Perseus’s bulging pelvis (and minuscule sex organs added by a later hand), however, fails to convey the confident, albeit neutered, appearances of the Gemini on folio 56v (plate 5) and of Aquarius on folio 59v (plate 11). The head of the Gorgon, Medusa, in Perseus’s left hand resembles the figure of Virgo on folio 56r (plate 4). It appears that the master painter provided the harsh-brown contour lines and then permitted an assistant to finish the work on folio 59r:

Perseus has bright stars in each shoulder, one bright one in the right hand, three stars in the left hand holding the head of the Gorgon, one star in the stomach, one bright star on the right side, one in the knee, one bright star in the foot, one bright star in the left thigh, two stars are in the shin, and three bright stars are around the head of the Gorgon. The head and the scimitar, however, are accepted as having one star each.¹²⁸

Lyra, also known as Fidis, has one star in each playing quill, one at the tip of both sides, one in both of the arm supports, one star in the base, and one star in the back. They total eight.¹²⁹

Cygnus, the swan, has one bright star in its head, five per wing, one star in the body, and one star in the tail. They total thirteen.¹³⁰

The first painter resumed work on folio 59v. The form of Aquarius (plate 11) resembles the figural style seen in the depiction of the Gemini on folio 56v (plate 5). There is a particularly strong resemblance between the picture of Aquarius and the image of the left-hand musical twin, Amphion. The description of Aquarius notes the water flowing from his overturned pitcher, which was his identifying feature: “Aquarius has two dim stars in his head, one bright star per shoulder, one bright star in the left elbow, one star in the right elbow and one in that hand, one per pectoral muscle, one in the right leg, and one bright star per foot. They total twelve. The profusion of water was also designated by thirty stars, from which two in the set are brighter than the rest, while the others are dim.”¹³¹

The representation of Aquarius in the Leiden Aratea on folio 48v portrays Ganymede, the cupbearer to the Olympian gods, with a Phrygian cap that recalls his Trojan heritage (fig. 50).¹³² The related image of Ganymede-Aquarius in Drogo’s copy of the Handbook of 809 portrays the mantle of the figure as a full cape slung over his left arm, opposite the overturned pitcher raised in his right hand. It is interesting to conjecture that the removal of pants (or chausses) from the half-clad Aratean image in Leiden (fig. 50) resulted in, or contributed to, the sexless physiognomy of the miniature in Drogo’s book. A similar disrobing of the clothed figure of Hercules in the Leiden Aratea also resulted in an arguably sexless heroic figure in Drogo’s copy of the Handbook of 809 (plate 2 and fig. 51). In the depictions of the Gemini found both in the Leiden Aratea and in Drogo’s manuscript (Madrid 3307), however, the figures are effectively nude (plate 5 and fig. 46). The first master painter in Drogo’s copy of the Handbook of 809 nevertheless depicted the Gemini without any recognizable indications of sex. Perhaps this demonstrated a sacralization of the figures that undermines the pure anatomical beauty of the human form, resulting in asexual, and occasionally sterile, nonclassical depictions. The depiction of Aquarius from the Recensio interpolata in Saint-Gall 902 on page 92 displays the figure fully clothed with a mantle (fig. 40). Ironically, the reversal of the left-right and back-front orientation of the depiction of Perseus in Leiden on folio 40v results in another sexless figure for the handbook in Madrid (plate 10).

FIGURE 50

Aquarius. Leiden Aratea, Bibliotheek der Rijksuniversiteit, MS Voss. lat. Q. 79, fol. 48v. Photo courtesy of Universiteitsbibliotheek Leiden.

FIGURE 51

Hercules. Leiden Aratea, Bibliotheek der Rijksuniversiteit, MS Voss. lat. Q. 79, fol. 6v. Photo courtesy of Universiteitsbibliotheek Leiden.

These changes argue in favor of the following conclusions:

1.  The artists of Drogo’s copy of the Handbook of 809 censored their representations of figural forms in the pictorial cycle by removing sex organs from their paintings of the constellations.¹³³

2.  The imagery in the lost Handbook of 809 harks back to pictorial prototypes related to those that also inform the creation of the cycle of miniatures in the Leiden Aratea for some, but not all, of the images, revealing a polycyclic approach to the creation of an astronomical pictorial program.

It is interesting in this context that the images of feminine constellations such as Virgo, Cassiopeia, and Andromeda in Drogo’s copy of the Handbook of 809 (presumably representative of the original archetypal cycle) are all fully clothed, unlike the bare-breasted image of Andromeda from the Leiden Aratea. The neutered presentations of the male constellations preserve, arguably, a more direct link to the classical pictorial content than do the feminine star pictures. Certain details, such as gender, were deleted in keeping with a contemporary Carolingian understanding of what constituted an appropriate and normative depiction of the stellar form, which overlapped with concerns about moral issues for the local monastic readership of the Handbook of 809.

One of Charlemagne’s favorite books was Augustine’s City of God, according to his biographer, Einhard.¹³⁴ Augustine was quite clear about the importance of genitalia and human sex for the human body after the Resurrection, in fulfillment of the promise of eternal life upon which the early Christian Church had established its greatest hope. In City of God, Augustine argues not only for the presence of both male and female sex within resurrected and perfected bodies but also for the disappearance of “sexual lust, which is the cause of shame.”¹³⁵ In fact, in defense of female genitalia, Augustine explained: “Now a woman’s sex is not a defect; it is natural. And in the resurrection it will be free of the necessity of intercourse and childbirth. However, the female organs will not subserve their former use; they will be part of a new beauty, which will not excite the lust of the beholder—there will be no lust in that life—but will arouse the praises of God for his wisdom and compassion, in that he not only created out of nothing but freed from corruption that which he had created.”¹³⁶ The perpetuation of sexual difference into the afterlife prolonged a celebration of God’s creation. This belief about the glorification of human bodies in the heavens was not extended to the images of the constellations rendered in human form. For the early medieval world, these images of heavenly bodies came closest to capturing the classical content that the Handbook of 809 sought to convey, and the androgyne constellations proclaim more than the necessity of astronomical study. Classicizing images of the constellations also emerged as the harbingers of Greco-Roman content in its purest form for later medieval and Renaissance reinterpretation. In the hands of Drogo and his pupils at Metz, his copy of the Handbook of 809 continued to be read by mortal men who occupied a world in which “the lust of the beholder” could be elicited. This resulted in presentations of the constellations denuded of their sex, which nevertheless preserved aspects of their gendered sexual and virile identities for early medieval readers.

Interestingly, a zodiac wheel associated with the series of manuscripts of the Recensio interpolata—in other words, manuscripts of the Revised Aratus, such as the one on page 100 of Saint-Gall 902 (fig. 52)—displays Aries with a laurel at the head of the new year, the two Pisces joined by a cord at their mouths, and Aquarius-Ganymede cloaked solely in a mantle gathered about his shoulders. One more conclusion can be advanced at this time: the representations of the zodiac signs in Drogo’s copy of the Handbook of 809 appear to refer back to standard images derived from zodiac wheels, but they do not always conform to any known cycle in particular. Certainly, no such cycle has survived. The Gemini twins in the zodiac wheel from Saint-Gall, for example, are clothed and bear arms. The more typically truncated depiction of Taurus was rendered full-length in Drogo’s copy of the Handbook of 809, as another counterexample to the archetypal Handbook of 809’s direct reliance on the zodiac linked to the Recensio interpolata in Saint-Gall 902 (fig. 52). In contrast, two of the remaining seven representations of the zodiac from Drogo’s book, Virgo (plate 4) and Sagittarius (plate 12), shared significant iconographic details with the zodiac on page 100 in Saint-Gall 902 (fig. 52). The zodiac sign Libra was not included in the Handbook of 809 or its catalogue of the constellations. The constellation, however, received its own depiction in the twelve-part zodiac wheel as a matter of course. This also could help explain, art-historically speaking, why the scales are separated from the angelic vision of Virgo/Dike in Drogo’s book, as they also are in the zodiac wheel. Similarly, the artists of the zodiac wheel in Saint-Gall 902 represented Sagittarius in the less common form of a satyr, as it is portrayed in Drogo’s handbook as well (and discussed below).

Reforming the Zodiac and Star Pictures South of the Ecliptic

The work of the first painter’s team breaks off in Drogo’s copy of the Handbook of 809 in the middle of folio 59v. The next six miniatures are the craftwork of a superior early medieval artist who preferred a large contour line juxtaposed with fine textural details. This second master painter rendered the forms of the constellations through fine hatching and feathery touches with great textural variety. These miniatures, despite their sorry state of conservation at present in Drogo’s copy of the Handbook of 809, are among the major accomplishments of the pictorial program. The style used to render these forms resulted from the artistic innovations of the team working in Metz. Their textual descriptions follow (plates 11–13):

Capricorn has one star per horn, one bright star in its nose, two stars in the head, one star beneath the neck, two stars in the chest, one star in the first foot, one star at the tip of the foot, seven stars in the back, five stars in the stomach, and two bright stars in the tail. They total twenty-four.¹³⁷

Sagittarius has two stars in its head, two stars at the tip of the arrow, one star in the right elbow, one star in the hand, one bright star in the stomach, two in the back, one in the tail, one in the first knee, one star at the tip of the foot, one star at the back knee. They total thirteen.

Aquila, the eagle, has four stars; of these the one in the middle is bright. The arrow has four stars: one in the tip, one in the middle, two at the other end.

Delphinus, the dolphin, has one star in its ear, two in its horn, three in the little pectoral fin, one star in the back, two stars in the tail. Altogether they total nine.¹³⁸

FIGURE 52

Zodiac. Stiftsbibliothek St. Gallen, Cod. Sang. 902, page 100. Photo: Stiftsbibliothek St. Gallen.

Orion has three bright stars in his head, a bright star in each shoulder, one dim star in the right elbow, one star in the right hand, three stars in the belt, three bright stars in the dagger, one bright star per knee, one star in each foot. They total seventeen.¹³⁹

Canis Major, the dog, has one bright star in its tongue, which is called Sirius, the Dog Star, one dim star in each of its front legs, two stars in the chest, three stars in the front paw, two in the stomach, one star in the left thigh, one bright star in the paw farthest back, and four stars in the tail. They total sixteen.¹⁴⁰

Beginning with Orion (plate 13)—which straddles the celestial equator, thereby rendering the constellation nearly universally observable¹⁴¹—and following the order established by Aratus, the list of the constellations shifts to descriptions of the constellations to the south of the ecliptic: “but many others rise below this, between the south and the track of the sun.”¹⁴²

Two important shifts take place in the dramatic presentations of these constellations. First, the more delicate technique of modeling the forms with three shades of the same color placed alongside one another was abandoned in favor of a dramatic contour and textural build-up of color with hatching, employed by a new master painter’s team. Second, this technique complements virtuosic moments of freehand painting unparalleled by other Carolingian artists working in the first half of the ninth century. For example, the body of the mythical hybrid half-goat, half–sea monster Capricorn (plate 11) displays harsh blue-grey contour lines and an exceptionally refined and confident, sinuous watery line washed across the folio surface to define the other side of the creature. Strident forelegs emerge through the freehand colored brushstrokes, without any discernible contour line beyond the knee joints, in confident, rapid washes. The hind legs of Canis Major show a dramatic contrast (plate 13). Different shades create a sense of overlapping perspective in the dog’s anatomy. If the first painter and his assistant employed a more controlled, quiet style motivated by economy and accuracy, then this second painter and his pupil were the dramatic expressionists at work within the small team employed at Metz.

Rapid brushstrokes and juxtaposed layers of color define the tortuous forms of Capricorn’s horns. Similarly, the hirsute pelt of the satyr Sagittarius (plate 12) defies the controlled contour line of the inner thigh on the right leg, by freely overflowing that bestial boundary with dark brown and tan ticks of paint. This chaotic display, reminiscent of manuscript illumination from Reims (discussed in chapter 3), nevertheless permitted the Carolingian painter to model the thigh’s cylindrical form.¹⁴³

The fan-like tails of the dolphin (plate 12) and Capricorn (plate 11) required the juxtaposition of a lighter and darker color to create a feathery effect. The forms recall the style displayed by the fan-like tails found in carved ivories of the Carolingian period made in Metz. This second master created the plumage of Aquila the eagle (plate 12) with a tan underpainting accentuated by medium brown and dark brown tick marks. Like the other images by this painter, the contour line was used to define one side of the bird’s body. The expressive potential of the layered hues (salmon-pink flesh tone for the talons, and brown hues ranging from ochre to chocolate) suggests the remainder of the fowl’s plumage with a sense of arrested motion, as though the bird of prey has just alighted on its fletcher’s perch, the Sagitta (or arrow). The talons themselves are fascinating from a codicological point of view, since the second master painter used the salmon-pink ink, at hand for rubrication in many Carolingian astronomical manuscripts, to render rapidly a gestural expression of their gnarled, knuckled form.

Although Orion (plate 13) has been badly damaged, the remaining paint best reveals the working technique of the second master painter. Light-blue washes define the left-hand side of the figure, opposite the darker contour lines on the right side of the legs. In some areas, such as the elevated right arm, light and thin contour lines were probably intended to define the form, yet would have been concealed by the introduced pigments. In other zones, such as the calf and shin of the hunter’s right leg, the washes supply the graceful curves of Orion’s musculature. Additional hatching identifies the elevated contour of the knee on the same leg. Applications of an even darker blue to the right forearm and Orion’s underbelly supply deeper shadows. More saturated red-orange lines define the folds of the mantle. Lighter washes of the same hue define its form.

The image of Sagittarius (plate 12) in the form of a satyr derives from an iconography associated with ancient celestial globes (figs. 53–55) and Carolingian zodiac wheels (fig. 52). Nearly all the important Aratean images of Sagittarius depict the hunter as a centaur. Sagittarius is portrayed as a centaur in the Leiden Aratea, for example (fol. 52v; fig. 21).¹⁴⁴ Aratus also suggests that he intended his Archer to be a centaur, since Aratus notes the “forefeet” of the constellation and consequently draws attention to the form of a quadruped.¹⁴⁵ This form is arguably the most ancient, with precursors in the ancient Near East. The satyr type could also have two heads (in which case one resembles a sparrow hawk). Egyptian sources gave rise to this alternative. A fourth, rarer possibility is the presentation of Sagittarius as a childlike creature.¹⁴⁶

As noted above, a consistent effort was made to provide information in the De ordine, which drew attention to the sorts of meteorological phenomena described in the Phaenomena. The south winds rose with violent force and bitter cold whenever “the sun meets up with Capricorn.”¹⁴⁷ Before the trying month linked to Capricorn, Aratus adds that “when the sun inflames the bow and the Drawer of the bow, you should put ashore in the evening and not continue to trust the night. A sign of that season and that month will be the rising of the Scorpion at the end of night. The Archer actually draws his great bow near the sting.”¹⁴⁸ This passage adequately describes the appearance of the lower, right-hand corner of the zodiac wheel from the Saint-Gall 902 copy of the Recensio interpolata on page 100 (fig. 52), mentioned above. As already noted, there was an apparent tendency for the artists of the Handbook of 809 to adopt aspects of the image cycle linked to the traditions of the Revised Aratus, revealing a preference for those manuscripts which included the text of the Recensio interpolata, as revealed by Drogo’s copy of the Handbook of 809. Whether the compilers of the imagery for the Handbook of 809 adopted the satyr type for Sagittarius because of a pictorial tradition linked to certain redactions of the Revised Aratus is impossible to say for certain. The drawing of Sagittarius on page 93 of Saint-Gall 902 relied on the more ordinary form of a centaur instead. There was not even consistency within any one manuscript’s pictorial program, which indicates just how tentative any hypotheses must remain about the archetypal pictorial program in the Handbook of 809.

FIGURE 53

Martin Folkes, illustration of the Atlante Farnese for Bentley’s Manilius of 1739, M. Manilii Astronomicon, PA6500 .M4 1739. Photo: Special Collections, University of Virginia, Charlottesville, Virginia.

FIGURE 54

Farnese Atlas Globe. National Archaeological Museum of Naples, inv. 6374. Photo © Vanni Archive / Art Resource, New York.

FIGURE 55

Farnese Atlas Globe, detail. National Archaeological Museum of Naples, inv. 6374. Photo © Vanni Archive / Art Resource, New York.

Given the widespread iconographic diversity revealed by the manuscript record, it is reasonable to conclude that a typically Carolingian polycyclic strategy of image compilation was at work in these astronomical codices.¹⁴⁹ Decisions about the sorts of pictures to include in various pictorial programs required the consultation of multiple sources either because of resource scarcity or because of resource availability. There is no reason to discount the latter possibility. The manuscript evidence does not support the conclusion that monastic scribes and illuminators, though in relative isolation, engaged exclusively in a straightforward rote practice of copying individual exemplary codices without reflection, taking one classical exemplar and restoring its tarnished splendor. In the Handbook of 809 and the traditions of the Revised Aratus manuscripts, the reoccurrence of iconographically related and clearly derivative forms of star pictures, discovered in tandem with illuminations displaying moments of aesthetic interpretation and painterly invention, suggests instead that a collaborative courtly workshop practice was fostered throughout the Frankish lands. Scribes and illuminators shared their visual presentations of the heavens with one another from Reims to Metz to Aachen, as the promulgated texts revealed the mysteries to be interpreted through artistry and were captured anew, time and again, on the parchment preserve of the foliated text. Both pictures and texts disseminated slippery traditions that migrated from one context to another without excessive discretion, so long as the image type remained, along with its text, an acceptable record of period scientia. For this reason it is best to recognize the general point that the satyr type of Sagittarius arose more commonly in conjunction with zodiac wheels or images related to the scant evidence pertaining to celestial globes (a point discussed in detail in chapter 3).¹⁵⁰ Either the satyr type or the centaur type of Sagittarius picture may have reminded readers of the Aratean admonishment to bring vessels into harbor each night during late fall and winter voyages. Incidentally, Aquila is also linked to navigational difficulties, whenever the eagle “rises from the sea at the departure of night.”¹⁵¹

The satyr form of Sagittarius in Drogo’s copy of the Handbook of 809 vied with the more ancient representational form of the constellation as a centaur. This relatively innovative break with other pictorial cycles containing Carolingian star pictures has suggested several important conclusions. First, the representations of the constellations within zodiac wheels or pictures related to star globes provided alternative forms of images for artists to choose to include in their visual representations of the heavens. Material objects, and not just manuscripts, were probably consulted when creating manuscript cycles of illustration. In the case of Carolingian star pictures, the material objects of choice were precisely those whose use facilitated mastery of the pedagogical aspects of the Handbook of 809, such as a celestial globe.¹⁵²

Second, the selection of a less common iconographic form, such as the satyr for Sagittarius, was a significant and purposeful decision made by a medieval painter. Such atypical forms nevertheless remained connected to the normative, iconographic meanings associated with the more common symbols.

Third, it is possible that the form changed for practical reasons. In fact, the Catasterisms 28 derived from Eratosthenes, as explained above, describes Sagittarius as a satyr. Douglas Kidd has argued that the satyr form of Sagittarius could more readily be differentiated from a suitably rendered star picture of the Centaur in astronomical texts and extended pictorial cycles.¹⁵³ In any case, Georg Thiele has noted that the satyr form of Sagittarius was also linked to Krotos of Sositheos in the texts of Hyginus.¹⁵⁴

The next three miniatures bear the mark of a style that is related to that of the second master painter but inferior in its expressive force. The images on folio 61r (plate 14) are the work of the assistant to the second master. Like the depiction of Canis Major (plate 13), the haunches and rear legs of the hare (plate 14) contrast dramatically with each other by the juxtaposition of the farther haunch cast in shadow and a lighter rear leg that comes forward in a form of overlapping perspective from the viewer’s vantage point. The graceful and sinuous form of the dog, which the second master rendered in a bold and confident profile, contrasts with the sketchy and mottled hues, which were intended to model the hare’s pelt. The colors of the hare’s coat resemble the basic color scheme of Leo on folio 57r (plate 6). This demonstrates that the first and second painters and their assistants all occasionally shared certain workshop practices. In fact, the vibrant hues rendered in gentle washes belong to a color scheme and style of painting associated with the manuscripts in Metz, such as the Drogo Sacramentary (fig. 33). This is an important result, for it permits a synchronic discussion in chapter 3 of the style of Carolingian manuscript painting in Metz.

There is no need in Drogo’s copy of the Handbook of 809 to posit two discrete campaigns of illustration or competing aesthetic impulses. The art-historical evidence argues instead in favor of four talented painters of sophisticated specialized training who plied their trade in medieval Metz for and alongside Bishop Drogo:

The hare (Lepus) has stars in each eye, two stars in its body, one star at the end of the tail, and one star per paw in the back legs. They total seven.¹⁵⁵

The ship, which is called Argo among the Greeks, has four stars in the stern, five on the side, three at the apex of the mast, and five stars below in the keel. They total seventeen, because the whole ship in the sky is not illustrated, but it is fashioned from the rudder to the mast.

Coetus has two bright stars in its tail, and five stars from the tail to the bend in its body; six stars are beneath its stomach. They total thirteen.¹⁵⁶

Interestingly, the assistant to the second painter created an innovative depiction of the vessel Argo (plate 14), which transported Jason and the Argonauts, according to legend. He chose to paint the ship in its entirety rather than follow the text and portray only the stern and midsection in Drogo’s book. Other Aratean examples, such as the miniature in the Leiden Aratea on folio 64v, adhere more accurately to the textual tradition, offering a bifurcated nautical vessel. When half of the Argo was dashed upon the rocks (the Symplegades), it rose like a fallen hero from the cold emptiness of the sea to assume its commemorative place in the celestial sphere. The image in Drogo’s copy of the Handbook of 809 is an original composition derived from other half-length boats of the Aratus recensions and a competing tradition of seaworthy nautical vessels.¹⁵⁷ The Carolingian assistant to the second painter completed the vessel with three bowsprits.

This assistant used the same technique as his master to supply a minimally modeled form, by placing a lighter tone next to a darker shade of the same or a related hue. For example, the bold contour line along the bottom of the boat dramatically defines its keel, using the same technique as the second painter in his stark rendering of Capricorn’s underbelly (plate 11). The lightblue and cinnabar of the sails recall the garments of Orion (plate 13), painted by his master. In palette and design, the link between master and pupil remains apparent, though the unrefined quality of the assistant’s work ensured that his participation was restricted to merely two folios.

The same holds true for the image of the sea monster Coetus (plate 14). According to the myth, Perseus rescued the sacrificial daughter Andromeda by slaying this sea monster.¹⁵⁸ As noted above, the tails of the dolphin (plate 12) and of Capricorn (plate 11) by the second painter display a sophisticated, feathery texture achieved through the overlapping of pigments and small, quick brushstrokes. His assistant painted the tail of Coetus with a geometric sterility instead. The feathery texture of the fins on the head of the sea monster resembles the work by the second master, and may have been used as an example to train his student. In any case, the assistant could not replicate the technique in the large wing-like fins below the belly of Coetus. The stark contrast between the evergreen contour and the lime-green body of the serpent visually separates into clumsy green stripes that do not coalesce into a presentation of rounded form. The technique was inspired by his teacher’s use of a stark contour line in an image such as that of Capricorn. The second painter’s assistant attempted to duplicate the pointillist technique that the first painter used to great advantage in his representation of the scaly flesh of Serpens on folio 54v (plate 1). In the image of Coetus, however, the dots merely provide a decorative pattern along the winding back and belly of the sea monster.

The painting of the second master displays a greater array of colors than those of the first artist. He also delicately renders the forms, which are painted freehand in certain areas but always delineated with a confident draughtsman’s sense of line, or ductus. The next four miniatures are also by the second master painter:

The river, which is called Heridanus, has three stars at the first bend, three at the second bend, and seven at the third bend, which are called collectively the Mouths of the Nile; these total thirteen. A star nearby is called Canopus.

Piscis Magnus, the Great Fish, has twelve stars placed in order from head to tail.

The sacrificial altar (Ara), or the shrine, has four stars: two stars are in the fire which burns where it was placed on the altar and two stars are in the pedestal (plate 15).¹⁵⁹

The Centaur has one star in each shoulder, two stars per elbow, four stars in the chest, ten stars in the rest of the body, two stars in the feet, two in the garment, two stars in the [Bacchic] thyrsus, which he carries, ten stars in the animal, which he holds in his hand, two stars in that hand, and five stars in the head. They total forty-three.¹⁶⁰

These four images on folios 61v–62r (plates 15–16) are among the most accomplished of the pictorial program, which indicates that the second master painter received the honor of carrying out the more complicated compositions in the manuscript.

It is important to recall that in Drogo’s copy of the Handbook of 809, Sagittarius and Capricorn were also painted by the second master painter. On these grounds, it is fair to propose that the second master painter was the most important master working in Metz. The second master painter was probably also the head of the artistic team of four painters encountered throughout this detailed analysis of the pictorial program. There was one assistant under the direction of each of the master painters, whose styles have already been discussed. The very fact that this team was working at an episcopal center on a scientific manuscript suggests that the artists may not have been permanently situated in Metz. Apparently, an astronomical manuscript was a specialized commission made at a particular site, such as Metz, for a specific recipient, such as Drogo. There is no evidence to suggest instead that a specialized scientific scriptorium arose in the ninth century to produce on speculation luxurious or functional schoolbooks, akin to the way that Touronian pandects emerged as a lucrative venture linked to the renowned monastic scriptorium of Saint Martin’s, beginning with the abbacy of Alcuin (d. 804) and continuing well into the tenure of lay abbot Vivian (d. 851).¹⁶¹ And the diversity of styles manifested by the individual copies of the Handbook of 809 under consideration confirms this result.

Given all of the evidence considered in part 1 of this volume, from the Annales Prumienses to this review of the pictorial program, it is highly likely that Louis the Pious was the royal patron who commissioned Drogo’s book and had the order carried out in medieval Metz, where his half-brother could oversee the project. How long a band of four painters was assembled in Metz is difficult to determine, given the limited number of illustrations in a copy of the Handbook of 809 and the specialized nature of the manuscript. It is not difficult to imagine a situation in which one or both of the two master painters were brought in from elsewhere for a unique commission. Following the creation of Drogo’s book of star pictures, they could have remained in Metz or, as itinerant painters, moved on to subsequent commissions throughout the Frankish lands. This theme is taken up again in chapter 3.

For now, it is important to underscore that these painters need never have worked together as a team on any other book. Related activity can be tracked during the ninth century. A group of four painters, however, coalesced around the specific project of creating Drogo’s book primarily because Drogo was a prelate with royal kinship ties to Charlemagne. But, as time would reveal, on the witness of his sacramentary and the gospel books from early medieval Metz (figs. 30–34), Drogo was also enamored with luxury codices. If the book is treated as an important fact, as argued in the introduction—that is, if the manuscript becomes a nodal point of creative activity in a Carolingian network of scribal composition and painterly creation—then the fluidity of forms and artistic styles entering into its creation can better be appreciated. In other words, against standard historiographic models that taxonomically cluster stylistic currents into schools (in the work of Koehler and Mütherich, for example), the evidence available in Drogo’s copy of the Handbook of 809 argues instead for a more dynamic scenario. Painters such as the second master found themselves in novel temporary environments that influenced the production of specific books. Those working conditions established limitations that informed the iconographic models available, and that resulted in diverse dynamic relationships among masters, assistants, and patrons. Pragmatic considerations of economy, the successful completion of the commission, and sites where standing libraries, such as Corbie, were available for consultation set limits on the degree of fluid exchange. To treat these four painters in Metz as a mobile atelier is, however, to repeat the methodological error identified here. On the contrary, the work of the painters in Metz on Drogo’s book offers insights into creative activity circa 830 in the Austrasian heartland, as four discrete early medieval lives converged. The book’s relationship to other manuscripts documenting eighth- and ninth-century creative activity is the focus of chapter 3, including a discussion of the innovative diagrams in the Handbook of 809 and the relationship of Drogo’s copy to other illuminated manuscripts.

Bold, stark contour lines separate the drapery of Heridanus into well-lit and shaded areas (plate 15). The second painter also used this technique in his attempt at modeling the form of Capricorn (fol. 59v; plate 11). The personification of the god for the River Po, otherwise known as Heridanus,¹⁶² is damaged and demonstrates the second painter’s lack of skill at rendering human faces. Large round eyes suggest sockets sketchily. The second master painter employed a large shadowy swath for visual effect. This vied, however, with his effort to render a reasonably accurate human form. In the profile of Sagittarius the satyr (fol. 60r; plate 12), and in the animal forms such as Canis Major (fol. 60v; plate 13), the second master painter used this technique to more satisfactory effect. Harsh contour lines divide out the regions of pectoral and abdominal musculature in the depictions of Sagittarius (plate 12), Heridanus (plate 15), and the Centaur (plate 16). The chubby fingers and small hands in these three miniatures also suggest that they are by the same artist who painted Orion (plate 13). These stumpy hands differ from the long outstretched fingers associated with the first master painter in his depictions of Cepheus, Cassiopeia, and Andromeda (plates 7–8) on folios 57v and 58r. The harsh contour line beneath the dorsal fins of Piscis Magnus (plate 15) delineates the upper boundary of the fish’s body and recalls how the same painter painted the Capricorn picture (fol. 59v; plate 11).

The signature textural studies of the second master painter supply an expert method for painting the transition from equine fur to human flesh in the image of the Centaur. A noteworthy novelty was the introduction of the white paint, added for highlights in both the image of the Sagittarius (around the arm drawing the bow in plate 12) and that of the Centaur (along its hindquarters in plate 16). Finally, many of the animal or hybrid forms painted by this second master display a ferocious or earnest eye: Capricorn (plate 11), Aquila and even Delphinus (plate 12), Canis Major (plate 13), and Piscis Magnus (plate 15).

The odd inclusion of a thyrsus, or shaft of pine cones, in the image of the Centaur (plate 16) appears to reinforce the distinction between the working techniques and philosophies of the two master painters in Metz. The thyrsus was a symbol of Dionysus carried by his followers.¹⁶³ But the thyrsus is mentioned in the text. A change in the artistic vision of the second compared to the first master painter fails to explain its inclusion. The second master painter created an acceptable interpretation of the hegemonic text of the De ordine.

Rather than dismiss the reference to the thyrsus as an insignificant detail or an odd authorial choice, it would be better to propose an art-historical rationale for its inclusion in the text. Any proposal should respect the prevailing tendency in the handbook to attribute oddities to scientific rather than mythopoetic origins. The textual source for the passage on the Centaur from the De ordine is the Basel Scholia. The term “veste” from the description of the constellation in Drogo’s copy of the Handbook of 809 on folio 62r appears in the other important Carolingian copies of the handbook as well (i.e., Reg. lat. 309, folio 98v, and Vat. lat. 645, folio 65r). This emendation was intended to clarify the confusing text of the Basel Scholia in Corbie. The Basel Scholia includes a list of the stars for the constellation named Chiron the Centaur. The passage on folio 38r–v of the Basel Scholia states that “in vestis ex [sic] in thyrso iiii” (in the garment [slung over the arm] and in the thyrsus are four stars).

The Carolingian editor of the handbook preferred to emend the words “in vestis” to the singular ablative form “in veste.” This would have been fine had the draughtsman who worked on the archetypal Handbook of 809 also included the garment slung over Chiron’s left arm in the image from the Basel Scholia on folio 38v. The copies of the Handbook of 809, such as Monza f–9/176, also leave any sort of garment out of the picture on folio 69r.¹⁶⁴ The emendation in Corbie was a sensible change to the text accompanying the sorts of pictures seen in Aratean manuscripts, because the Centaur typically did have only one animal slung over his shoulder (typically a panther) that in certain manuscripts resembled a free-flowing pelt suggestive of a mantle. That motif also resembled a skin or pelt in the depiction of the Centaur once found in the Leiden Aratea, if a later copy from Boulogne-sur-Mer of the missing miniature can be trusted (Bibliothèque municipale, MS 188, folio 29r, ca. 1000).¹⁶⁵

Given the textual traces that lead back from the De ordine to the Basel Scholia, it is reasonable to identify the Centaur depicted in Drogo’s book as Chiron, the pious teacher of Achilles. This centaur was reportedly an expert on medicine and astronomy. In Etymologies IV.ix.12, however, Isidore of Seville extended the Centaur’s résumé to include expertise in veterinary science.¹⁶⁶ According to Isidore, Chiron’s treatment of animals resulted in his representational form as a hybrid creature. In the image from Drogo’s copy of the Handbook of 809 (plate 16), the rabbit (perhaps a hare) replaces the sacrificial panther and the thyrsus replaces the club found in the image from the Boulogne-sur-Mer Aratea. In other words, the choice of a centaur holding a thyrsus and rabbit for display was intentional. The image of Chiron paid homage to classical medicine, which was intrinsically linked to astronomy and even astrology in both the classical world and in the Middle Ages. Unlike the mantle, the cone-tipped rod of the thyrsus remained in the pictorial tradition because of the demands placed upon the picture, which differed from the demands placed upon the text. The thyrsus drew attention to the mythological physician, who is pictured similarly in the image accompanying the copy of the Basel Scholia. The second master painter, like the first, apparently felt compelled to accentuate these details, which permitted the iconographic connections to medieval science; the pagan, spiritual role of the rod as a reference to Dionysian bacchanals had been suppressed. Indeed, some might say that the second master painter in Metz painted the thyrsus in Drogo’s book in such a way that it resembled a hunting spear used for slaying the rabbit, even though the rod was referred to explicitly as a thyrsus in the text.¹⁶⁷ Hunting was much more palatable to ninth-century royals than were pagan rituals.

The Centaur’s name was also removed from the text of the De ordine. The erasure shows the effort of the editors in Corbie to accentuate scientific points while suppressing certain mythological details. In this instance, the silences are as informative as the ideas expressed in the text. Details pertaining to the form of the constellation in space needed to be retained. Further artistic decisions were motivated by a spirit of scientific pragmatism, provided such iconographic decisions conformed to, or at least did not vie with, Christian orthodoxy.

The image of the Centaur in the Cologne Aratus on folio 166v is probably related to the one in the archetypal Handbook of 809 and was probably its immediate temporal precursor within the Frankish lands (ca. 805 as opposed to ca. 809). As suggested above, the Revised Aratus manuscript from Cologne provided a model for the creators of the original Handbook of 809 and the painters of Drogo’s copy to bear in mind. The Centaur from the copy of the Basel Scholia is also related to the form that probably harks back to a prototypical depiction of Chiron. It was the text of the Basel Scholia, however, that designated the Centaur as Chiron. This detailed analysis of the text and image of the Centaur in Drogo’s copy of the Handbook of 809 provides art-historical evidence for the direct reliance of the Handbook of 809 text about the constellations, the De ordine, on the text of the Basel Scholia.

The last four depictions in the handbook were all done by the assistant to the second master painter:

Serpens, the Serpent, which they call the Hydra, has three bright stars in its head, six stars where it first bends, three stars at the second bend, two stars in the fourth bend, and eight bright stars from the fifth bend through the tail. They total twenty-six.¹⁶⁸

Corvus, the raven, is on the tail [of the Hydra] looking out over a depression [in the serpent’s body]; the raven has one star in its eye, two stars in its wings, two in its tail feathers, and one star per foot. They total seven.¹⁶⁹

The Crater, or the urn, which was placed after the first bend of the serpent’s body, has two dim stars at its lip, three stars in the midsection, and two stars in the foot. They total seven.

Anticanis, Canis Minor, has three stars.¹⁷⁰

The assistant to the second master revealed his true expertise in the second representation of the Raven on folio 62v (plate 17). The image is first displayed on the back of the Hydra on folio 62r (plate 16). In both, the attention to detail is limited to the heavy contours likewise encountered in the images on folio 61r, for example (plate 14). The assistant to the second master painter created the solid form of the second raven freehand and without underdrawing, carefully reticulating the gnarled appearance of the bird’s talons. The image of the raven is an impressive example of medieval pen-and-ink drawing.

The Hydra on folio 62r is a lackluster attempt at modeling the physiognomy of the creature (plate 16). The assistant to the second master similarly exaggerated the juxtaposed layers of hues in the image of Coetus on folio 61r (plate 14). Neither image captures the textured modeling seen in the image of Serpens on folio 54v (plate 1), attributed to the first master painter. The image of the Crater (plate 16) on the back of the Hydra resembles the second image on folio 62v (plate 17). Both have a nearly hemisphericaltonal arrangement, such that the right side displays an exaggerated shaded zone typical of this painter’s style. In the individual image of the Crater, the assistant to the second master painter mimicked his teacher’s preferred technique of hatch marks texturing quasi-modeled volumes. The hatching, however, became a disappointing series of banded striations in the hands of the pupil.

The image of Canis Minor (plate 17) clearly shows the contrast in style between the assistant to the second painter and his master. Whereas the depiction of Canis Major on folio 60v (plate 13) shows the stern gaze typical of the second hand, the assistant inserted a beady pupil into the eye socket of Canis Minor on folio 62v (plate 17). The second master painter painted the far back leg of Canis Major darker than the one closest to the observer, whereas the assistant misinterpreted the style and divided both rear legs of Canis Minor into striped zones. In Canis Minor the left rear leg has the dark shaded strip on the bottom, whereas the right rear leg closest to the viewer has the shaded region on top, along the dog’s haunches, demonstrating the second assistant’s confusion. This assistant was clearly learning from his teacher but had yet to master the logic of his master’s techniques.

As noted already, the repetition of the images of the Raven and the Crater was a conscious decision motivated by the preference for visual parataxis in the Handbook of 809. This form of parataxis was a departure from earlier Aratean pictorial programs and a creative presentation of the pictorial program attached to the De ordine in Drogo’s copy of the Handbook of 809. As a result, the Carolingian astronomers and artists of the handbook, as well as their later copyists, demonstrated a value placed upon the individual representations of celestial forms for scientific understanding and also for the aesthetic pleasure of their learned patrons.¹⁷¹

Greco-Roman Visions of the Stars and Their Christian Avatars: Panofsky, Saxl, and Carolingian Creativity

This iconographic analysis has revealed that the creative process in Metz was much more than a straightforward copying or reclamation effort. The tendency to render the individual images of the constellations according to an early medieval pictorial strategy of visual parataxis indicates that even longstanding forms could be recombined in new ways. The probable reliance of aspects of the archetypal Handbook of 809 on textual and visual components of the Revised Aratus also places the cycle of illustration in Drogo’s personal copy within an interpretative framework that emphasizes active processes of organization (ordinare) and critique (emendare).

Adopting a compositional strategy appropriate to encyclopedic literature, the Revised Aratus of the late eighth century included excerpts of texts that recast the traditional significance of Aratus’s Phaenomena. The selection of texts from authors such as Isidore of Seville did more than pay homage to the legacy of classical antiquity, as Le Bourdellès would suggest.¹⁷² Isidore was not only an encyclopedic exegete of classically inspired natural history and the liberal arts, including astronomy, but imbued the rich rhetorical, literary, and scientific legacies of his pagan and Christian auctores (such as Varro, Pliny, Macrobius, Martianus Capella, Donatus, Solinus, Servius, Vergil, Lactantius, Cassiodorus, Jerome, and Augustine)¹⁷³ with a novel early medieval Christian auctoritas. Nowhere was this point made more explicit or with more visual force than in the fifth wheel diagram (Munich, Bayerische Staatsbibliothek, Clm 16128, folio 16r; fig. 1) depicting the mundus-annus-homo, or the microcosmic-macrocosmic harmony, in Isidore’s De natura rerum.¹⁷⁴

As addressed above in the discussion of the image of Hercules on folio 55r, the first master painter included pictorial references to the Nemean lion and club (plate 2). But what is missing from the image of Hercules is arguably the most important factor to appreciate. Unlike the image of Hercules in the Cologne Aratus on folio 156v (fig. 8), in the image of Hercules in Drogo’s book, the mythological reference to the Tree of the Hesperides has been ignored. This reflects a point of view advanced by Isidore in book III of his Etymologies: “Therefore, observations of the stars, or horoscopes, or other superstitions that attach themselves to the study of the stars, that is, for the sake of knowing the fates—these are undoubtedly contrary to our faith, and ought to be so completely ignored by Christians that it seems that they have not been written about.”¹⁷⁵

As Isidore had prescribed, one of the express responsibilities of a compiler or early medieval author was to relay what could be helpfully gleaned from the classical authors while packaging the astronomical information in terms that did not border on heterodoxy and while effacing questionable content. Such deliberate selection to guide the reader through a proper mode of interpretation consists of a creative process of textual composition and image collection that is referred to here as exegetical emendation. The methodological consequences of this design process are discussed in greater detail in chapter 4. For now, it suffices to note that the interpretative or exegetical strategy of compilation appropriate to the creation of the Handbook of 809—which derived in part from the same pedagogical impulse that Isidore had manifested in his Etymologies—recognized that the ultimate utility of training in the liberal arts lay in the application of such knowledge by professional clerics. As Arno Borst has argued, the Handbook of 809 benefited from this kind of interdisciplinary spiritual and tactical focus (see the introduction).¹⁷⁶

The constellation Hercules from Drogo’s copy of the Handbook of 809 provides an example of this focus. The image of Hercules himself was essential to the star catalogue contained within the De ordine. The mythological content pertaining to Hercules’s attack on the serpent that guarded the golden apples in the Tree of the Hesperides was omitted, even though this larger, elaborate image was linked to the Revised Aratus and depicted in the Cologne Aratus (plate 2 and fig. 8). Similarly, on folio 14r, the Basel Scholia reports, “Hic est Hercules sup- anguem incumbens clava” (Here is Hercules, attacking a serpent with a club), before the corresponding image of the constellation follows on folio 14v (fig. 7).¹⁷⁷ The image from the copy of the Revised Aratus in Cologne was created in 805, and the text associated with the Basel Scholia shows Hercules attacking the serpent, which makes it all the more remarkable that the image in Drogo’s copy lacks the same pictorial references.¹⁷⁸ This shows that early medieval processes of image selection for large natural-historical cycles of illustration involved exegetical strategies of discovery. Choosing certain iconographic types and disregarding others when culling through extant prototypes is a meaningful rather than rote activity, and deviating from standard options to advance a non-mythological point of view in the Carolingian Handbook of 809 was an act of purposeful emendation on the part of Adalhard’s team.

This level of creative involvement in the production of new manuscript anthologies has not traditionally been associated with aspects of the Carolingian renewal. It is easier to focus on the Carolingian star pictures that remain and use them as evidence of both the extant and the lost pictorial traditions that would have been passed down from late antiquity to Frankish artists. This was the important art-historical project of Fritz Saxl and Erwin Panofsky, as mentioned in the introduction.

Saxl and Panofsky’s “Classical Mythology in Mediaeval Art” appeared in 1933 in Metropolitan Museum Studies.¹⁷⁹ In this essay, the coauthors argued that the Franks had played a vital role in linking the High Renaissance celebration of classical form and proportion to their ancient forebears. Along the way, however, Carolingian artists necessarily undermined the pure expression of classical Latinitas to advance a period-specific agenda and Holy Christian imperial ideology. However, Panofsky and Saxl attest to an unnecessary putative divide, in which they claim that the effort to promote a nascent yet already hegemonic Christian interpretation of the star pictures obviates the possibility for a reinsertion or continuity of classical content into the Carolingian images of the heavens. In their words, “The Carolingian Renaissance differed from the ‘Rinascimento’ of the fifteenth century in many respects. Where the latter was based on the irresistible feeling of the whole people and was brought forth in popular political and spiritual excitement, the earlier was the result of the deliberate efforts of a few distinguished men, and thus was not so much a ‘revival’ as a series of improvements in art, literature, calligraphy, administration, etc.”¹⁸⁰

Panofsky and Saxl are here willing to recognize that Carolingian star pictures were closer to the classical models than other, later images of the constellations “in style and technique,” thus adhering to long-standing “representational traditions” offering “integral unities of subject matter and form.”¹⁸¹ This measured praise for the ability of Frankish artists and scribes to copy with rote precision came bundled with a disparaging dismissal of Carolingian creativity, disregarding the possibility of exegetical emendation or the role of conscious decisions in the formation of the text and image cycles contained within manuscripts such as the Handbook of 809. It is not truly fair to say that the Frankish artists and scribes working for Adalhard copied either the Aratean texts or aspects of their corresponding cycles of illustrations when compiling the original Handbook of 809. On the contrary, original strategies of selection and composition resulted in the organic creation of a novel Carolingian presentation of the cosmos that laid the foundation for similar decisions in Drogo’s later copy.

The more definitive statement of Panofsky’s belief in the exceptional renewal of classical Latinitas during the High Renaissance appeared in his 1960 Renaissance and Renascences in Western Art.¹⁸² Here Panofsky altered the emphasis from an inquiry into the transmogrification of celestial forms over the longue durée to a heartfelt apology for the celebrated uniqueness of the classical renewal experienced in fifteenth-century Florence and Rome.¹⁸³ Carolingian star pictures, in Panofsky’s view, could only represent aspects of classical form and content through the sedulous copying efforts exerted by the Frankish painters.

It is important to underscore here a vital point and thereby to correct an occasional misconception. Panofsky celebrates the Carolingian painters who created classically inspired images of the zodiac and constellations, especially when this required novel pictorial research reflecting upon examples of cycles in mosaics, on wall paintings, on celestial globes, or in extant manuscripts that were believed to be classical and not just early Christian precedents. Carolingian artists saved from oblivion these classical prototypes and breathed new life into these “figures (or groups of figures) classical not only in form—this, needless to say, applies to numberless motifs handed down by the Antique to Early Christian art and thus held in readiness for the Carolingian revival—but also in significance.”¹⁸⁴ Panofsky develops his argument: “Carolingian art, we remember, revived and reemployed scores of ‘images’ in which classical form was happily united with classical content. . . . The iconographical significance of these images, however, remained unchanged.”¹⁸⁵ And again: “Nowhere in Carolingian art, however, do we seem to encounter an effort to infuse into a given classical image a meaning other than that with which it had been invested from the outset. . . . We are confronted with quotations or paraphrases, however skillful and spirited, rather than with reinterpretations.”¹⁸⁶ It is in this context that Panofsky moots his “Principle of Disjunction”: if a medieval artist adopts a classical form for a composition, then its corresponding content is typically recast in compliance with Christian orthodoxy; alternatively, if a medieval artist opts to illustrate a narrative with its roots in pagan antiquity, the corresponding visualization of the content will reflect spatio-temporally contingent medieval mores and courtly life.¹⁸⁷

Panofsky could not envision, however, the precise situation in which the creators of the Handbook of 809 found themselves. Under Adalhard, an official delegation recombined and revised myriad fragments of classical text and images, assembling the truthful record of Carolingian astronomy one sanctioned piece at a time in the years that followed the synod of 809. The text of the De ordine was illustrated with classical forms and advanced classical content, but did so within a programmatic Christian platform calling for religious reforms.¹⁸⁸

The agenda pragmatically set the parameters on the semiotics of interpretation and established the framework in which the creation of the Handbook of 809 could take place. In other words, the exegetical emendation of the text-image pair pertaining to Hercules in the De ordine preserved the classical Aratean form and content of the constellation within the star catalogue, while it simultaneously excluded the pagan mythological narrative alluded to by the Tree of the Hesperides from the folios of Drogo’s copy of the Handbook of 809.

Modern readers and viewers may at times glimpse the late antique classical past through the shadowy reserve of the choices made by the Carolingian scribes and artists working for Adalhard in Aachen. But to focus exclusively on the images in Drogo’s copy of the Handbook of 809 as records of lost pictorial precedents (or as records at one level of remove from the archetypal Handbook of 809) is to miss a vital insight about Frankish life—namely, that the artistic and creative decisions which resulted in the extant pictorial programs in the handbook supply firsthand information about Carolingian decisionmaking processes. The extant star pictures do not simply inform the art historian about iconographic precedent. Rather, these pictures (such as the one of Hercules) declaim aspects of the Frankish identity of their creators, revealing facts about Frankish priorities, values, preferences, and prejudices, without the need for a Weltanschauung.