The organ was the King of Instruments long before Mozart coined the phrase in a letter to his father. Prized for its multifarious sounds and technical complexity, the instrument has been used to celebrate powerful rulers and religious institutions since its invention in the third century BCE. An organist was crowned victor of the music festival at Delphi c.90 BCE, and the Emperor Nero seems to have been more interested in new organ designs than in restraining barbarian revolts. Pippin's court was furnished with an organ as a gift from the Byzantine Emperor in 757, and in 826, Charlemagne's son, Louis the Pious, ordered that one be built for his palace at Aix. The organ's splendour was ideally suited to liturgical use, and by the thirteenth century Aegidius of Zamora describes the organ as the only instrument used by the Church, an association which it still carries in most parts of the world today. Much of the music recorded here, or improvisations in a similar style, would have been suited to the ample acoustics of a sacred setting. The two fifteenth-century Mass settings, played in alternation with Gregorian chant, present the earliest preserved organ music for liturgical use as it might have been performed in European cathedrals almost six hundred years ago. But surviving sources also include keyboard arrangements of secular songs and dances, some of which attest to a high degree of musical sophistication among the clerics and courtiers who cultivated organ-playing during the fourteenth and fifteenth centuries.

A major problem in performing this repertoire today is the choice of instrument. No one organ can possibly reflect the nuances of tuning and timbre suggested by the different styles of late-medieval keyboard music. A Pythagorean tuning based on pure fifths is suggested by the prevalence of that interval in music of the Robertsbridge Codex, while the significant placement of simultaneous thirds in some pieces of the Buxheim Orgelbuch benefits from a meantone temperament. Delicate chanson intabulations are well suited to the single registers of a chamber organ, while some of the drone-based pieces benefit from a full organ sound in a large acoustic. How to make all of this live again for the late twentieth-century listener? I decided to search for this variety on one organ, the modern reconstruction of a fifteenth-century instrument that stands in a swallow's nest gallery on the south wall above the rood screen in Basel's Predigerkirche. There was an instrument in this position during the first half of the fifteenth century, but details concerning specification survive from an organ built in 1487 by Johannes Tugy, so this is the one upon which the new organ is modelled. Bernhardt Edskes, a pioneer in the field of organ archaeology, researched the issues leading to the design of the organ, which was built by Sebastian F. Blank of Herwijnen, Netherlands. The main manual of the organ is a large principal chorus whose registers can be played together as a Blockwerk or divided to isolate individual principal and mixture timbres. The pedal (FF-c) is permanently coupled to the main manual (F-g''a'') extending the compass an octave lower, as is documented on several late-medieval instruments. The notes F-c can thus be played by either the manual or pedals. Although the original Tugy organ had one manual only, there was enough room for the addition of a new Brustwerk, which contains a flute, gemshorn and regal. This makes it possible to imitate the sounds of small chamber instruments, as well as to have a second manual for repertoire requiring this. (There is evidence for secondary manuals on organs as early as 1386.) The mechanical bellows system is constructed so that it can be pumped by hand, creating nuances in the winding that impart some lovely musical effects. (I am very indebted to my diligent Kalkanten who sacrificed two evenings of sleep for the flexible winding heard here!) These resources offer many possibilities for the performance of the repertoire included in this recording. The wonderful acoustic of the church provides a rich environment for the sound, allowing it to blend and reverberate as it must have in the fourteenth and fifteenth centuries.

Having determined which instrument to use, there remained the problem of programming this unusual and diverse repertoire, which is often composed of very short pieces. I have tried to give modern listeners a broad context within which to appreciate each work by grouping them in interesting ways. Where possible, both sacred and secular music from each source is presented together. The Kyrie-Gloria movements that begin and end the recording are performed in alternatim with Gregorian chant to re-create the sense of rhythmic ebb and flow and the contrast between fluid vocal lines and highly embellished instrumental writing. These large structures frame excerpts illustrating the variety of early keyboard music. Following the first Kyrie-Gloria are two other pieces from the Faenza Codex: the repetitive rhythmic figures of Bel fiore dança are rendered clearly on a light flute, while the message of the chanson "Or sus, vous dormés trop" ("Now awake, you sleep too much") is emphasised by the use of the organ's "birdsong" device to accompany one of the refrains. (We encountered a great deal of authentic birdsong when recording this music in the wee hours of the morning, so you might hear it in other places as well.)

Keyboard settings of dances and vocal music are also found in the Robertsbridge Codex, the oldest surviving keyboard music, which may have been composed around 1359, while the French court was held prisoner in England following the Battle of Poitiers. Retrové is based on a medieval dance known as the estampie, which features repeated endings. The open ending leads back to a repeat of each section of the piece, while the closed ending terminates each section after it is repeated. The title Retrové may refer to the performer's need after each new section of the piece to "find again" a way into the musical text of the endings, which serve as a sort of refrain to the new material presented in each section. The ingenious integration of repetition is the hallmark of this work, which constitutes a fascinating example of late-medieval minimalism.

In addition to estampies, the Robertsbridge Codex includes intabulations, or arrangements, of two motets by Philippe de Vitry from the Roman de Fauvel, an early fourteenth-century musico-drama satirizing the pompous society of the French court. The first of these, recorded here, is based on Firmissime/ Adesto/Alleluya Benedictus. The three voices of the motet rarely appear together, so the piece is generally in two parts, with a decorative upper line featuring triplet patterns above the longer note values of the lower voice. The performance of motets on the organ may antedate the earliest musical sources, for Juan Ruiz de Hita's Libro del Buen Amor, written c. 1330, indicates that motets and chansons are played on the organ, as found in the later Robertsbridge motets and Faenza songs.

More song intabulations are found in a manuscript folio included among surviving fragments of a fourteenth-century musical treatise found in the binding of Groningen, University Library incunabulum no. 70, a fifteenth-century printed book. The folio of music contains arrangements of two French songs, Asperance and Empris domoyrs, where the countertenor was omitted and the cantus was elaborated over longer note values in the tenor. This two-part texture, with a florid right-hand melody moving over slower notes in the left hand, is characteristic of the late-medieval music heard so far.

The rest of the recording is devoted to musical sources from the area of present-day Germany, where three-part writing and the use of pedals are more frequent. One of the earliest extant sources specifying performance with the pedals is the tablature of Adam Ileborgh, dated 1448. The manuscript contains 5 Praeambula and 3 elaborations of a German song melody, known as Mensurae (from the Latin for "measure") because they divide the tenor notes into different numbers of beats--three, two, and six respectively. The praeambula are extremely short pieces intended to establish a mode, and they are used here to introduce the mensurae. The headings for some of these preludes stipulate the use of manuals only or with pedal. The designation manualiterfor the pieces with faster tenors suggests that the drone-like sustained tones of the other style were intended to be played by the pedals, as demonstrated in Praeambulum 4. Thus interpreted, the Ileborgh tablature constitutes the earliest documented use of double pedaling in the organ repertoire.

A short fifteenth-century treatise of German origin entitled De musica arte, now in the Munich Bayerische Staatsbibliothek,contains a discussion of ways to create music over a tenor, providing valuable insight into the process of composition/improvisation. The anonymous author of the treatise includes a list of possible figures for the discantus called "tactus," each composed of four notes lying within the intervals of a second, third or fourth. They are grouped in three categories: "ascendentes" (rising), "descendentes" (falling), and "indifferentes" (staggered patterns). The secret of making good music lies in adjusting these patterns to suit the tenor over which they sound, as shown by the untitled piece appended to the treatise, whose tenor may be based on a German song.

The most important collection of the fifteenth century is the Buxheimer Orgelbuch , a compilation of over 260 pieces, including song intabulations, cantus-firmus compositions, free preludes, and exercises in composition. (Because of their didactic nature, none of the studies in this latter group are included here.) The first set of selections from Buxheim combines Redeuntes, pieces based on a repeated tenor note, with Praeambula. Both of these genres may have served liturgically as intonations, and they display originality in the use of figurative patterns and textural change. In the Redeuntes, the reiterated tenor, played by either the left hand or pedal, creates a drone over which the discantus weaves various figures, especially of the sort described in the Munich treatise (see above). This structure is ideal for extemporization, and undoubtedly such pieces were often improvised, the reverberant acoustics of medieval churches accentuating the crescendo effects produced by the repetitive texture.

The praeambula demonstrate great melodic and harmonic creativity.Like those in the Ileborgh tablature, these praeambula are freely invented pieces that exploit different keyboard textures. The Buxheim pieces are generally longer and more developed, however, with binary and ternary structures featuring solo melismatic lines, homophonic passages in three parts, and polyphonic sections with one or more independent voices. The Praeambulum super mi opens and closes with a melismatic line, where the occasional intrusion of a group of five and six notes in a passage of regular four-note groups provides lyrical freedom, creating an improvisatory quality in the execution of the melodic line. (This same type of rhythmic irregularity is used in the discantus voice of the Ileborgh Preambula and the following Mensurae.) By contrast, the discantus voice in the central section of this Buxheim Preambulum suggests a more regular pulse to highlight the dotted rhythms and four-note patterns of figuration. The Praeambulum super f is also in three main sections, a solo melismatic line leading into a succession of three-part chords, followed by dense writing juxtaposing duple and triple rhythms. The Praeambulum super C and the Redeuntes in C employ many of the same figurations in the discantus voice. Yet compared to the unvarying tenor of the Redeuntes, the Praeambulum exploits textural change to a far greater degree, moving suddenly from the melismatic solo to three-part writing that is homophonic in places, to two voices where the tenor adds sonority and rhythmic articulation. Despite the brevity of most of the Buxheim Preambula, these pieces exploit the polyphonic versatility of the keyboard, where voices are easily added or taken away, creating crescendo and diminuendo effects on the organ without recourse to stop changes or enclosed expressive pipe divisions. A variety of dynamic and textural change distinguishes the preambula from the formulaic style exhibited in much late-medieval organ music.

The simpler Buxheim Preludia serve as introductions to some of the song arrangements in the next section of the recording. Je loe amours is based on a ballade by Gilles Binchois, one of seven settings included in the manuscript. The rondeau that served as a model for Adieu mes tres belle is also by Binchois, attesting to the great popularity of his vocal music. The beautifully ornamented notes of the top voice are ideally suited to the intimacy of a chamber organ, which I have tried to evoke with the delicate 4' flute. The composer of J'ay pris amours has not been identified, but this was a widely disseminated rondeau. The countertenor is largely absent from the second half of the piece, creating a reduction to two-part texture. Portugaler is based on an anonymous chanson for which the text has not survived. My choice of a simple 8' Principal registration is meant to accentuate the beauty and seeming nostalgia of the ornamented cantus.

The repertoire in the Buxheimer Orgelbuch may be seen as the culmination of the move from two- to three-part polyphonic writing for the organ by the second half of the fifteenth century. This was effected by developing the countertenor into an increasingly independent part. In the Ileborgh Tablature, the countertenor is often a doubling at the third or fifth of sustained tenor notes. In the Buxheimer Orgelbuch, the countertenor is written into the score and exhibits an independence that transcends its doubling function in earlier music. It is enlivened rhythmically and covers a large range, frequently crossing underneath the tenor. A good example showing how the intabulator changed the model is found in the Buxheim intabulation of Dufay's ballade, Se la face ay pale. Dufay's original countertenor is discarded, and the different third voice added by the intabulator creates harmonic changes that greatly alter the sound of the piece. The tenor line of the original chanson serves as little more than a framework for the relentless figuration in the discantus.

The Kyrie and Gloria Beata Maria Virgine are movements based on the plainsong for Mass IX, Cum jubilo. The first Kyrie and the Christe are exceptional because they are written consistently in four voices, the pedal indications corresponding to the lowest voice throughout. The vibrant majesty of these settings corresponds to the shimmering fullness of the organ's principal chorus, capped by the high-pitched Zimbel mixture. This block of sound (known as a Blockwerk before separable registers made it possible to isolate the ranks) is the characteristic timbre of the late-medieval organ, and it must have filled with awe those listeners fortunate enough to witness its splendour. The alternation of powerful organ tone with human voices might be considered an allegory for the hierarchy inherent in the medieval worldview, with the omnipotent God reigning over the terrestrial estates. Although the centuries that divide us from this music have also distanced us from the theocentric philosophy that led to its conception, perhaps this musical re-creation can convey to the modern listener something of its original magnificence.