A show tune, a sixty-two-year-old man and a number one

On the Billboard chart dated 9 May 1964, a record by a sixty-two-year-old trumpeter reached number one in the United States and ended a fourteen-week occupation of that position by the Beatles, who had held it with three different songs in sequence. The record was Hello, Dolly!, a show tune from a Broadway musical that had opened that January, cut the previous December for the Kapp label by a man who was reported to have needed persuading to record it at all. Nothing about the event fits the standard story of Louis Armstrong greatest recordings, which is why this account of them begins there. The standard story says that the essential Armstrong records were made in Chicago in the 1920s, that the trumpet playing on them remade jazz, and that everything after was commerce, clowning and gradual decline. The chart of May 1964 is a documented fact that story cannot digest, because the same singer and the same trumpet that the canon seals inside 1928 were, almost forty years later, the most commercially powerful force in American popular music for one week in the spring. Either the technique had survived the decades intact, or a sixty-two-year-old had outsold the Beatles on charm alone. This article works through the sessions in order and argues the first of those.

The claim defended across the thirty-four sections that follow is precise. The catalogue divides by market, by format and by employer, but it does not divide by technique. The lead playing on the records of 1925 to 1928 and the lead playing on the records of 1954 to 1964 are the same craft applied to different material: a trumpet part built as a paraphrase of the melody that grows bolder with each chorus, a sense of time that sits the beat where the whole band can lean on it, a vocal delivery that treats a lyric as a rhythm part, and a solo architecture of statement, development and climax compressed into whatever space the format allowed. What changed between the periods was the space. A three-minute race-catalogue side, a dance-band arrangement, a twelve-inch LP track and a two-minute pop single each rationed that architecture differently, and the records differ because the rationing differed. Judged by what the playing does rather than by what the market asked, the line from the Hot Five to Hello, Dolly! is continuous, and the famous break in the middle of the catalogue is a feature of the criticism rather than of the records.

That argument has to be earned session by session, and the evidence is unusually good. Armstrong recorded for forty-eight years, across every format the industry used, and the documentation is deep: session ledgers, discographies assembled over decades, and scholarship by writers who disagree with each other about almost everything except the dates. Dan Morgenstern, the discographer and critic who annotated more of these sessions than any other writer, spent a career insisting that the records be judged decade by decade rather than waved past. Ricky Riccardi, the archivist whose research at the Louis Armstrong House Museum corrected personnel listings that reference books had repeated for years, represents the documentary end of the argument. Gary Giddins, whose book and film under the name Satchmo made the continuity case to a general audience, represents the critical end. Their work is used throughout what follows, and where they disagree with the older literature, the disagreement is named rather than smoothed away.

A note on method before the sessions begin. Every record discussed here is identified by the facts that locate it: the title, the label, the session date, the personnel as the best discographies give them. Where those discographies dispute a chair or a date, the dispute is reported as a dispute, because a catalogue article that prints contested personnel as settled fact teaches the reader to trust the wrong thing. No lyric is quoted anywhere in this article, and no solo is transcribed. The records are described by what can be heard and verified: who played, what was played, how the performance is built, what it sold, and what the reviewers and the reissue programmes did with it afterwards.

How the catalogue is usually divided

Open almost any short account of Armstrong and the catalogue arrives pre-divided into two unequal halves. In the first half, running roughly from 1923 to 1929, a young cornetist from New Orleans records with King Oliver, accompanies the great blues singers, leads the Hot Five and Hot Seven studio groups, and in a handful of Chicago sessions in 1928 produces records, West End Blues above all, that are treated as founding documents of jazz as a soloist’s art. In the second half, running from 1930 to his death in 1971, the same man fronts big bands, sings popular songs, appears in films, forms the All Stars, records with Ella Fitzgerald, and scores pop hits in his sixties. The halves are not merely unequal in length. They are unequal in status. The first is called art and the second is called entertainment, and the adjectives usually arrive attached to a judgment about the playing itself: the early records are described as invention, the later ones as repetition, the early man as a revolutionary and the later one as an institution trading on the revolution.

The division structures everything written downstream of it. Anthologies of his work weight the 1920s so heavily that a newcomer could listen for hours without reaching 1930. Survey histories of jazz, including the division history told in the history of jazz and blues music, allot the young Armstrong a chapter and the older one a paragraph, because the chapter structure of jazz history itself, one style succeeding another, needs him young: he is the hinge between ensemble jazz and solo jazz, and a hinge has no biography after the door has swung. Even his admirers have often enforced the split. The revivalist critics of the 1940s valued the early sides as relics of a purer practice and treated the big band records as contamination. The bebop generation that followed had its own reasons to keep him in the past, since a music announcing itself as the future requires the previous era to look finished. Who decided which half of this catalogue counts? That question opens here, in section two, and it is answered in the next section, because the answer turns out to be datable: the division was not discovered in the records but drawn on top of them, at a specific time, by identifiable people, for reasons that had as much to do with what was in print as with what was on the discs.

Part of the division’s durability is that it flatters a general theory of artistic careers, in which a figure burns early and coasts late, and part is that the early records really are extraordinary. Nobody defending the later work needs to pretend that the Hot Five sides are ordinary, and this article does not. The 1928 sessions are as good as their reputation says. The question is narrower and more testable: whether the playing that produced them stopped, diminished or merely changed employment. That question can only be answered by listening across the whole catalogue with the same ears, which is what the remaining sections attempt. The division also had a commercial shadow. Through the LP era, the 1920s material was reissued in scholarly sets with annotated booklets, while the 1930s Decca big band records circulated in budget compilations with no documentation at all, so each generation of listeners met an early catalogue presented as history and a middle catalogue presented as product. Presentation is not a small force in how records are judged, and section thirty-two returns to what the reissue programmes did to the shape of his reputation.

For now, the division itself is the exhibit. It assigns the records to two careers, implies that one man made the first and another, lesser man made the second, and it survives because it organises forty-eight years of sessions into a story short enough to repeat. The next section takes the story apart and asks when it was built.

Why that division is a critical construction

The division of the catalogue into a canonical youth and a commercial old age was drawn by critics decades after the early records were made, not by the market that bought them and not by Armstrong, who went on recording the supposedly superseded music, in remakes and in concert, for the rest of his life. That is the first rung of this article’s ladder, and it closes the question opened in the previous section. The young Armstrong’s records sold inside a segregated catalogue system, the race series that Okeh, Columbia, Paramount and Gennett maintained for Black buyers, and within that system the blues accompaniments and the Hot Five sides were working commercial product, reviewed rarely and prized locally. The elevation of the 1920s sides into a canon began in the 1930s, in France as much as in America, when Hugues Panassié and the Hot Club de France treated the early records as the authentic article and began, in print, to rank the periods. Panassié’s admiration for the young Armstrong was genuine and informed, and his preference hardened into a framework: the hot early records were jazz, and the later popular records were a falling away from it. The framework travelled because it was written down, and what is written down first tends to govern what is written later.

The second force was generational. The bebop musicians of the 1940s defined their music against the swing era’s showmanship, and Armstrong, by then the most famous jazzman alive, became the convenient symbol of the older dispensation. Scott DeVeaux, the historian whose work on jazz historiography showed how the music’s history came to be written as a sequence of styles overthrowing each other, supplies the mechanism here: a history organised as succession needs each style’s founder to stop mattering once the next style arrives, and Armstrong’s refusal to stop, commercially or artistically, made him an embarrassment to the plot. The plot won in the textbooks. Meanwhile the records that would complicate it, the 1930s big band sides and the 1950s concept albums, were the hardest of his work to hear in good pressings with serious annotation, so the verdict circulated in the absence of the evidence. Morgenstern spent years making exactly that point in his liner notes: that the middle of the catalogue was condemned largely unheard, by critics working from a handful of reissues.

The third force was Armstrong’s own stage manner, and it belongs in this section because it distorted the reception of the records. Readings of his grin, his handkerchief and his stage persona as racial submission, readings made in print by writers who had access to the public performance but not to the private archive, attached a charge of accommodation to the later work that the records themselves do not support. Those readings are a subject in their own right, and the argument about them belongs to the companion debate article in this cluster; here it is enough to note the mechanism. A judgment about the man on stage was transferred to the man on record, and the transfer spared critics the labour of listening to forty years of sessions. When the private papers, the tapes he recorded at home and the correspondence became available to scholars late in the twentieth century, Thomas Brothers among them, the documentary picture that emerged was of a musician in deliberate control of his presentation from the beginning. The grin had been a professional instrument in 1928 too. It simply had not been charged against the early records, because the critics who loved those records encountered them after the fact, on disc, without the stage act attached.

So the division is a construction with a date, authors and uses. It was drawn in the 1930s, hardened in the 1940s, and preserved by reissue policy afterwards. None of that makes the early records less great. It makes the later records a question to be settled by listening rather than by inheritance, and the sessions table below lays out the territory the listening has to cover: forty-eight years, every format, one technique to test.

Session or album	Approximate date	Personnel	Market it was made for	What it does that the others do not
King Oliver’s Creole Jazz Band, Gennett sessions	April 1923	Oliver and Armstrong cornets, Johnny Dodds, Honoré Dutrey, Lil Hardin, Baby Dodds, Bill Johnson	Race catalogue 78s	First recorded Armstrong solos, inside a two-cornet ensemble
Bessie Smith sessions, Columbia	January and May 1925	Armstrong cornet, Fred Longshaw piano and organ	Race catalogue 78s	Accompaniment raised to a second voice behind the reigning blues singer
Hot Five, first session	November 1925	Armstrong, Lil Hardin Armstrong, Kid Ory, Johnny Dodds, Johnny St. Cyr	Race catalogue 78s	A studio group with no stage existence, built around his lead
Hot Seven sessions	May 1927	Hot Five with Pete Briggs tuba, Baby Dodds drums, John Thomas trombone	Race catalogue 78s	Added rhythm weight and the stop-time solo showcase
Sessions with Earl Hines	June to December 1928	Armstrong, Hines, Fred Robinson, Jimmy Strong, Mancy Carr, Zutty Singleton	Race catalogue 78s	Two virtuosos in dialogue; the ensemble becomes a frame for soloists
Weather Bird duet	December 1928	Armstrong trumpet, Earl Hines piano	Race catalogue 78s	A two-man record: the whole band reduced to a conversation
Ain’t Misbehavin’	July 1929	Armstrong with Luis Russell’s orchestra	Theatre-connected popular 78	The stage song carried by trumpet and voice in equal shares
Paris session	October 1934	Armstrong with a Paris-based orchestra	European market 78s	The big band format tested in front of the audience that canonised him
When the Saints Go Marching In, Decca	May 1938	Armstrong and his orchestra	General market 78s and jukeboxes	A sacred standard recut as swing for the broad popular audience
Town Hall concert	May 1947	Armstrong with a small group drawn from the period’s New York players	Concert recording, LP era approaching	The public relaunch of his small-group playing
Louis Armstrong Plays W. C. Handy	July 1954	All Stars: Trummy Young, Barney Bigard, Billy Kyle, Arvell Shaw, Barrett Deems	LP concept album	The blues songbook of his apprenticeship revisited at full length
Satch Plays Fats	Spring 1955	All Stars with Billy Kyle and Barrett Deems	LP concept album	The Waller songbook, including remakes that answer the 1929 originals
Ella and Louis	August 1956	Armstrong, Ella Fitzgerald, Oscar Peterson, Herb Ellis, Ray Brown, Buddy Rich	LP, Verve prestige catalogue	The duet album as an art form: two voices, one rhythm section, no orchestra
Hello, Dolly! single	December 1963	Armstrong and the All Stars	Pop single and Broadway tie-in	A show tune taken to number one by a sixty-two-year-old
What a Wonderful World	August 1967	Armstrong with orchestra and strings	Pop single	A ballad performance whose chart life arrived in two instalments, 1968 and 1988

The earliest sessions as a sideman

Armstrong’s first recordings were made in April 1923, in Richmond, Indiana, in the studio that the Starr Piano Company operated for its Gennett label, and he made them as the second cornet in somebody else’s band. King Oliver’s Creole Jazz Band had come up from New Orleans to Chicago, and Armstrong, who had idolised Oliver, had been summoned north the previous summer to join it. The Gennett sessions of 5 and 6 April 1923 produced the records on which his playing first survives: Chimes Blues, which carries his first recorded solo; Weather Bird Rag; Dipper Mouth Blues; Snake Rag; Canal Street Blues; Mandy Lee Blues and a scattering of others across those two days and a return visit later in the year. To understand what a sideman’s job was on such a date is to understand the conditions under which the whole early catalogue was made. The band recorded acoustically, gathered around a recording horn, with the quieter instruments placed forward and the loud ones held back, and the Gennett studio stood close enough to a railway line that passing trains could spoil a take. Nobody in the room was making art for posterity. They were making stock for a catalogue sold to Black buyers through furniture stores and mail order, at a pace of several titles a day.

The discipline shows in the playing. A second cornet in the Oliver band had a defined part: fill under the lead, support the harmony, take the breaks the arrangement assigned. The remarkable thing in the Gennett sides is how often the ear is drawn to the second chair. On Chimes Blues the brief solo he takes, sixteen bars of it, is already organised the way his mature solos are organised, as a melody that could be sung, stated plainly and then pressed harder. On Dipper Mouth Blues the featured solo is Oliver’s, three choruses played with a mute that became the record’s signature, but the two cornet breaks the two men play together, short harmonised figures worked out between them, became a set piece that bands copied for years. The sideman sessions of 1923 and 1924, for Gennett, Paramount, Okeh and Columbia as a hired cornet behind bandleaders and singers, are the apprenticeship on record: roughly two years in which he learned how a studio worked, how much playing three minutes could hold, and how to make an entrance register on a disc that gave him no visual help at all.

That last lesson matters for everything after. On a stage, Armstrong’s presence was physical and comic as well as musical. On a record, only the sound travelled, and the young sideman’s sound already travelled further than the men around him. Producers began asking for him by name. Clarence Williams, part pianist, part publisher, part fixer of the New York blues recording scene, used him repeatedly through 1924 and 1925, including on the Red Onion Jazz Babies sides with Sidney Bechet, where the two men spar across Cake Walking Babies with a freedom nothing else on record that year approaches. The personnel credit on that December 1924 Gennett session was argued over in early discographies, a useful early warning that even the foundational Armstrong records carry contested entries, a theme this article flags wherever it recurs. By the autumn of 1925 he was the most in-demand cornetist in the race catalogue business, and the demand had a structure: leaders hired him to lift ordinary sessions, singers hired him to answer them, and the records he touched tended to be the ones buyers remembered. The Gennett operation itself shaped what the sideman records could show. Richmond was a small Indiana city with a piano factory’s recording studio attached, reached by train, and the sessions there were day trips into primitive conditions: a small room, a recording horn, a band arranged by distance from the horn to balance its voices. Oliver’s men were veterans of the room by April 1923, and the records they made preserve an ensemble art at the moment before electrical recording would soften its edges. Armstrong’s place in that art was second cornet, the harmony chair, and the records show the harmony chair’s peculiar freedom, to decorate the spaces the lead leaves, to answer the lead’s phrases, to build the breaks that Chimes Blues made history with. The sideman years look, from the catalogue’s height, like a conservatory whose curriculum was the music’s whole foundation: ensemble discipline from Oliver, section discipline from Henderson, and the accompanist’s art from the singers, three educations running concurrently in 1923, 1924 and 1925, each audible in every record he made afterward.

His October 1924 session with Ma Rainey, cut for Paramount in Chicago weeks before the move to New York that Henderson’s offer required, makes a fitting hinge between the educations: the dean of the classic blues singers, a small group drawn from Henderson’s orbit playing in the older manner, and Armstrong, already New York bound, decorating See See Rider with the restraint that the singers’ records taught. Within a month he was in Henderson’s trumpet section. The catalogue’s first phase closes with the accompanist fully formed, which is why this article insists on beginning there: the soloist of 1925 was built by the sideman of 1923 and 1924, and the building blocks were other people’s records.

The accompaniments for blues singers

Between 1924 and 1926 Armstrong spent a large share of his studio time doing a job the record business considered routine: accompanying blues singers. The race catalogues sold singers, and a singer’s session needed a small band or a single horn behind the voice, supplying introductions, answers between the vocal phrases, and a final chorus to close. Armstrong played that role for Ma Rainey, for Clara Smith, for Trixie Smith, for Sippie Wallace, for Alberta Hunter, for Eva Taylor and for a line of others whose names now survive mainly in discographies. The October 1924 Paramount session behind Ma Rainey, which produced her See See Rider Blues, placed a twenty-three-year-old cornetist behind the most established blues voice in the business, and the record shows what the job could become in his hands. He does not decorate her phrases. He answers them, in her register, with a gravity that matches hers, and the record’s later reputation rests on that duet quality as much as on the singing.

The accompaniments deserve attention as training as much as the famous sessions do. A singer’s chorus is twelve bars of blues with the words carrying the meaning, and the horn behind it learns economy: when to enter, how long an answer can run before it crowds the next line, how to build a closing chorus that lifts the record without stealing it. Armstrong’s fills from these years are a study in that economy, and the phrasing he developed answering singers, the sense of where a breath falls and how a line wants to end, never left his playing. It is there in the trumpet solos of the 1950s and in the singing of the 1960s, because the accompanist’s discipline and the vocalist’s discipline are the same discipline. He also learned the catalogue from inside. The blues singers’ repertoire, the W. C. Handy compositions, the Fats Waller songs he would meet later in the theatre, formed the songbook he drew on for the rest of his career, and the two great songbook albums of the 1950s, examined in sections twenty-four and twenty-five, are this apprenticeship returning at full length and full price.

Why did the business use him this way? Partly availability and price, since a cornetist who could read a singer in one take saved studio money. Partly something producers could hear and buyers responded to. A record with his name nowhere on the label still moved when his horn was on it, and by 1925 the record men in Chicago and New York knew it. Gary Giddins has argued that these accompaniments, dozens of them, amount to a second, shadow catalogue running alongside the famous one, and that a listener who knows only the Hot Five knows the soloist but not the musician. The point stands up. The sideman records show the technique being built, chorus by chorus, in the service of other people’s songs, which is also a fair description of the pop records of his sixties. The market changed around him twice; the service did not.

Personnel on the blues dates is where this article’s rule about contested entries bites first. Most of the sessions are documented in label files and were reconstructed reliably by the discographers, but the early literature credited Armstrong with cornet work on sides where later research, Riccardi’s among it, has reassigned the chair, and a few sides long listed as his are now held to be the work of contemporaries whose sound resembled his. The honest summary is that the core of the accompaniment catalogue is secure and its edges are argued, and that the argument itself is evidence of how much anonymous session work he did: only a cornetist who was everywhere could be plausibly claimed for records he never made. The economics of the accompanist’s trade explain both its volume and its anonymity. Blues singers were recording properties with followings; their accompanists were session labour, paid a fee for the date and uncredited on the label, and a cornetist in demand could stack dates for competing labels in a single week under the loose exclusivity of the period. Armstrong’s fee income in 1924 and 1925 came substantially from this labour, and the labour’s discipline, learning a singer’s material quickly, in the studio, without rehearsal infrastructure, is part of what the records preserve. The race catalogue’s credits lagged the music badly, and part of the modern discographer’s work, Morgenstern’s and Riccardi’s generation systematising what the pioneers had sketched, has been the assignment of these anonymous choruses to their player by ear, session by session. The assignments carry margins of error, which this article has flagged where they touch its narrative, but the central fact is not in dispute: the young Armstrong is present, identifiable by the sound of his invention, across dozens of other artists’ records, and any complete edition of his work must count those records as his.

Trixie Smith’s Railroad Blues, cut for Paramount in March 1925 with a Henderson-derived group, shows the accompanist working a singer of the second rank up to her best record: the performance is hers, and the choruses behind and between her verses are constructed with the same care he gave Smith and Rainey, which is the professional’s definition he never abandoned. The habit of care, applied indifferently to masterpieces and programme fillers, is the trait that links the sideman years to the songbook albums of the 1950s, where the same conscientiousness would be lavished on Waller and Handy. Careers are usually described as rises; the catalogue describes this one better as a standard, set early and met, with remarkably few lapses, for forty-eight years.

Playing behind Bessie Smith

Did Armstrong record with Bessie Smith?

Yes. Armstrong accompanied Bessie Smith at two Columbia sessions in New York in 1925, in January and again in May, with Fred Longshaw at the piano and, on some titles, the organ. The January date produced the St. Louis Blues that stands among the most celebrated blues records ever made, and the pairing was not repeated after 1925.

The two sessions are small in number, fourteen titles across the year, and outsized in consequence. Bessie Smith in 1925 was the biggest-selling blues singer in the race catalogue, a Columbia artist whose records carried the label’s whole race series, and the pairing with the twenty-three-year-old Armstrong was a meeting of the catalogue’s reigning voice with its rising horn. What makes the records extraordinary, and what lands the second rung of this article’s ladder, is that the accompaniments for blues singers, Smith’s above all, are among the most instructive records Armstrong ever made, and they predate the work he is famous for. The method he would use for fifty years is fully formed here: listen to the singer, answer in the spaces, raise the temperature a degree at a time, and take the final chorus as a summation rather than a display.

St. Louis Blues, recorded on 14 January 1925, is the summit of the method. W. C. Handy’s composition is taken slowly, with Longshaw’s harmonium sustaining a church gravity behind Smith’s voice, and Armstrong’s cornet moving through the record like a second, wordless singer. He answers her lines so closely that the answers feel composed for them, though the record was made in the ordinary run of a commercial session. Later critics have treated the performance as a kind of summit meeting, and for once the superlative habit fits the evidence: the record was a strong seller in its day and has never been out of the anthologies since. The other titles from the two sessions, Reckless Blues, Sobbin’ Hearted Blues, Careless Love Blues, You’ve Been a Good Ole Wagon and the rest, vary the recipe without weakening it, and the May session loosens into something close to conversation between the two principals.

The pairing did not continue, for reasons that belong to the business rather than to any documented falling out: Armstrong’s price and profile rose, his Chicago base pulled him back west, and Columbia built Smith’s later sessions around other accompanists. Her own career, its Voice, its touring empire and its end, is told in the companion life of Bessie Smith, and this article deliberately leaves her story there. What belongs here is the residue in Armstrong’s catalogue. The Smith sessions taught the record-buying public to listen for the horn behind the voice, and they taught Armstrong the exact skill his later fame would monetise: making a song somebody else was singing sound as though it could belong to no one else. When he became the singer, in the 1930s and after, he accompanied himself with the same attentiveness he had given Smith, the trumpet answering the voice, the voice answering the trumpet, one musician holding both chairs of a duet he had first played in her records.

There is a coda worth recording. In 1954, when George Avakian produced the album of W. C. Handy songs discussed in section twenty-four, Armstrong was returning to the composer of St. Louis Blues after twenty-nine years, at the head of his own band, on the new long-playing format. The 1925 side is the apprentice piece for that album: same songbook, same accompanist’s intelligence, raised to the scale of a headline artist. Critics who treat the Handy album as a late peak disconnected from the early work miss the line that runs through both, and the line starts behind Bessie Smith. The texture of the Smith sessions deserves one closer look, because Longshaw’s instruments give the records their singular gravity. On some titles the piano is joined or replaced by a harmonium-like reed organ, whose sustained chords lay a churchly floor under the blues, and Armstrong’s cornet moves over that floor in long tones and short answers, a horn learning how much space a voice needs. The January date’s St. Louis Blues became the session’s monument, the record by which the singer’s art and the accompanist’s were jointly reintroduced to every later generation through its constant reissue, and its fame has somewhat eclipsed its companions: the May session’s I Ain’t Gonna Play No Second Fiddle, on which singer and cornet trade phrases in a marital spat set to music, shows the partnership at play, and Careless Love Blues, from the same spring dates, shows it in pure lament. Fourteen records is a small shelf, but the shelf holds the complete grammar of an art form: how a horn accompanies a voice. Everything Armstrong later did with his own voice, and everything he did beside Fitzgerald, is written in that grammar.

There is also the question of what the sessions did for him commercially, and a catalogue should answer it plainly. Smith was Columbia’s race catalogue’s flagship, and her records were the label’s prestige pressings in that market; to be chosen, at twenty-three, as the cornetist for her return to a small-group setting was a professional certification that travelled. The Henderson job had made him known to musicians; the Smith records made his sound known to the record-buying public of the race market, under another artist’s name, in the season before his own name went over a band. When Okeh titled the November 1925 session Louis Armstrong and His Hot Five, the name it was selling had been built in other people’s grooves, and the buyers of the Smith sides formed the first audience that followed it.

The Oliver band records

The Creole Jazz Band records of 1923, and the further Oliver sessions Armstrong played on into 1924, hold a place in the catalogue out of proportion to their number, because they preserve the sound he grew up inside and the mentor he measured himself against. King Oliver played cornet with a mute vocabulary Armstrong admired all his life, using plungers, cups and bottles to make the horn speak, growl and plead, and the younger man’s job in the band was to orbit that lead. The records are ensemble music first: collective improvisation in the New Orleans manner, two cornets, clarinet, trombone, piano, banjo and drums, with the melody carried by the front line as a braided single cloth. Yet the braid keeps parting around the second cornet. On Snake Rag the breaks flash between the two horns. On the celebrated Dipper Mouth Blues, Oliver takes the long muted solo that musicians memorised, but the record’s frame, the two-cornet break figures the pair had worked out together, is a dialogue between teacher and student so tight that early listeners sometimes could not tell where one man’s part ended and the other’s began.

That fusion was the band’s advertised marvel. The two-cornet breaks, short harmonised bursts played in unison rhythm, were rehearsed effects, Oliver devising them and Armstrong executing alongside him, and they announced that a dance band’s excitement could be composed as well as improvised. The Gennett sides preserve them best. The later Oliver records, for Paramount in the autumn of 1923 and for Okeh and Columbia in 1924 as the band’s personnel shifted and Armstrong’s role grew, show the student edging toward the microphone. On Riverside Blues, cut for Paramount in November 1923, his breaks have a swagger the Gennett sides only promise. Thomas Brothers, whose study of Armstrong’s New Orleans formation is the standard account of these years, reads the Oliver records as the sound of an apprentice who has already absorbed the lesson and is waiting for the chair to empty. Oliver’s band broke up in Chicago in 1924, the two men parted without a quarrel on record, and the records they left together became foundation texts for the revival generation, reissued and studied while their makers moved on to different music.

A word of discographical honesty belongs in this section too. The Oliver sessions are among the most re-researched in all of jazz, and the accepted personnel has shifted over the decades as Gennett’s files and musicians’ recollections were re-examined. Honoré Dutrey’s trombone, Bill Johnson’s banjo and bass work, Baby Dodds’s drums and Lil Hardin’s piano form the accepted core of the 1923 Gennett band. A handful of titles from the band’s later, looser sessions have had their cornet chairs debated, and a small number of records once assigned to the Creole Jazz Band have been moved to other groups entirely. This article prints the accepted lineups and notes the existence of the contested edges, which is the most the evidence allows and all a reader needs: the sound of the two cornets together is not in dispute, and it is the sound, not the ledger, that built the legend.

What did Armstrong take from these records into the rest of his career? Three things that never left him. A conviction that a band’s front line is a conversation, which survives in his dialogues with singers, pianists and trombonists for fifty years. A muted vocabulary learned from Oliver and deployed, sparingly and to devastating effect, whenever a record needed a change of voice. And the professional habit of the featured break, the eight or sixteen bars in which one player steps forward and settles the record’s argument, which becomes, in the Hot Five and after, the Armstrong solo itself. The Oliver records’ afterlife belongs in this section because it governed what later ears could hear. The Gennett sides are acoustically primitive, the balance horn-led, the bass nearly absent, and for two generations of listeners they arrived through reissues of varying care, their ensemble subtleties masked by surface noise. That Armstrong’s reputation nevertheless began there says something about the solos’ power to penetrate bad reproduction, and the modern restorations, which present the Gennett and Paramount masters with a clarity their makers never heard, have measurably shifted the critical picture toward the band and away from the soloist alone: heard clean, the records are ensemble achievements with a revolutionary second cornet inside them, rather than solos with accompaniment. The shift matters for this article’s thesis in a small way. It demonstrates, at the catalogue’s very origin, that judgements of Armstrong’s periods are hostages to the state of the records’ availability and reproduction, a condition that the decline argument, built during the worst decades of the big band records’ availability, never escaped.

Oliver’s own trajectory after 1923 closes the section with a melancholy symmetry. The band that made the Gennett records broke up within the year, Oliver’s later recording career faltered on bad teeth, bad management and worse luck, and he died in 1938 in Savannah, working outside music, while his second cornetist was the biggest Black bandleader in America. The catalogue keeps the proportion straight: Armstrong’s recorded remarks about Oliver, across forty years of interviews, are a pupil’s gratitude without qualification, and the Oliver sides remained in his own collection and his own esteem to the end. Whatever the canon disputes among the critics, the two men’s records divide into a master’s and a pupil’s who surpassed him, and both of them knew the order of events, even if only one of them lived to be asked about it often.

The Fletcher Henderson sides

In the autumn of 1924 Armstrong left Chicago for New York to join Fletcher Henderson’s orchestra at the Roseland Ballroom, the most prestigious dance band engagement available to a Black musician in America, and for roughly a year he recorded as the Henderson band’s hot cornet. The sides, made for Columbia, Vocalion, Paramount and other labels that leased the band, show a different Armstrong from the Oliver records: not a partner in a two-cornet braid but a featured soloist inside written arrangements, entering at marked places, taking his choruses against a full reed and brass section, and sitting down again. Don Redman, Henderson’s arranger, was building the grammar of big band jazz in those years, and Armstrong’s choruses taught him what a jazz solo could be asked to do. Redman began writing spaces for the soloist as structural events, and the modern big band arrangement, with its alternation of ensemble and featured improvisation, carries the mark of those sessions.

The Henderson sides that matter most are the ones where the soloist breaks the frame. On Copenhagen, recorded in the autumn of 1924, his choruses ride the arrangement with a rhythmic freedom the written parts around him do not attempt. On Sugar Foot Stomp, the band’s 1925 remake of Oliver’s Dipper Mouth Blues, the solo Oliver had owned is handed to the former sideman, and Armstrong plays the master’s set piece in the master’s city, faster, higher and with a tail of invention at the end that turns homage into annexation. Shanghai Shuffle and Money Blues show the same process at slower tempos, the cornet singing over the band’s figures. Musicians across New York heard these records and the Henderson band live, and the influence ran through the trumpet sections of every ambitious orchestra in the city. Coleman Hawkins, the band’s tenor saxophonist, was already a star; the records document the two men in the same room at the moment the soloist-centred design of jazz was being decided.

The Henderson year also completed his education in the business of recording. Henderson’s band recorded constantly, for competing labels under competing names, and Armstrong learned how a small stock of arrangements and a hot soloist could be sold across the whole catalogue system. When he left New York in November 1925, homesick for Chicago and for his wife Lil Hardin, who was building his solo career from there, he left as the most discussed young instrumentalist in Black music. The Henderson records had made him a soloist in fact. What they could not do was make him a leader, because the band’s name, the arrangements and the billing were Henderson’s. The next phase of the catalogue, the records that bear his own name, begins the moment he is back in Chicago, and it begins, improbably, with a band that did not exist. The Henderson records repay attention to what Armstrong did to the band’s own arrangements. Henderson’s book was written music, scored for sections, in the manner that Don Redman, the band’s arranger, was pioneering, and the soloist’s task inside it was to fill allocated bars. Armstrong filled them by treating the allocation as a stage: his choruses begin inside the arrangement’s harmony and depart from its rhythm so completely that the band sounds, for sixteen bars at a time, like his accompaniment rather than his employer. The practice is audible on The Stampede and Copenhagen, and it taught a generation of arrangers, Redman included, to write with the soloist’s departure in mind, which is one reason the Henderson sides belong in a catalogue of Armstrong’s records and not only in Henderson’s. His records with the band also include the odd vocal novelty and the sessions where he shares the trumpet section with Joe Smith, whose lyrical style beside Armstrong’s fire gives the 1925 band recordings a two-voice trumpet drama that discographies have long treasured.

The departure, when it came, was engineered rather than drifted into. Henderson’s organisation was losing money and engagements, the Roseland residency gave Armstrong a New York profile he had not arrived with, and Lil Hardin Armstrong, in Chicago, was pressing the case that the accompanist’s chair was now beneath him and that a job at the Dreamland Cafe, where her band played, waited with his name over it. He went home in the autumn of 1925 with the New York sides behind him, Copenhagen and The Stampede and the Clarence Williams sessions among them, and with a reputation in two cities and in the trade press of both. The Henderson chapter closes the apprenticeship: in fourteen months he had been the sideman of the leading dance orchestra in America, the accompanist of its greatest blues singer, and the featured guest of its most experimental small groups. What remained was to lead, and the leading began within weeks.

The Hot Five begins

What are the Hot Five recordings?

The Hot Five recordings are the sides Armstrong made for Okeh in Chicago between November 1925 and 1928, leading a studio group that never worked as a band in public: Lil Hardin Armstrong on piano, Kid Ory on trombone, Johnny Dodds on clarinet and Johnny St. Cyr on banjo, behind the cornet of the leader himself.

The first session took place on 12 November 1925. Armstrong was in Chicago playing theatre and cabaret jobs, in Erskine Tate’s orchestra at the Vendome and soon with Carroll Dickerson at the Sunset Cafe, and Okeh’s Chicago operation wanted records from the cornetist whose name had been rising through the Henderson and blues dates. The group assembled for the purpose was a studio construction, drawn from his New Orleans circle: Kid Ory, the trombonist whose band Armstrong had played in before leaving home; Johnny Dodds, the clarinetist from Oliver’s band; Johnny St. Cyr, a New Orleans banjoist and guitarist; and Lil Hardin Armstrong, whose role in the enterprise was larger than the piano chair suggests, since she organised, composed, rehearsed what little rehearsing was done, and understood, earlier than her husband did, that the records were a career and not a sideline. The three titles from that first day, My Heart, Yes I’m in the Barrel and Gut Bucket Blues, announce the format at once. On Gut Bucket Blues the leader introduces his players by name in a spoken roll call between the choruses, a piece of salesmanship for a band that had no stage on which to be introduced, and the playing itself is loose, funny and exact, all three qualities held together by the lead cornet.

A question opens in this section that the article will hold for seven more. What changed between the 1925 sessions and the 1928 ones? The records at the start of the Hot Five run and the records at its end are both canonical, both were made for the same label in the same city with overlapping personnel, and they do not sound like the same conception of what a jazz record is. In 1925 the cornet is first among equals in a collective improvisation. By the end of 1928, as section sixteen will show when the loop closes, the band has become a frame built around two soloists, and the music’s centre of gravity has moved permanently. The sessions in between, the Hot Seven expansion, the duet records and the arrival of Earl Hines, are the mechanism of that change, and they repay being followed in the order they happened.

Early Hot Five highlights came quickly. The February 1926 session produced Oriental Strut and Muskrat Ramble, the latter becoming a standard of the Dixieland repertoire ever after, and Heebie Jeebies, the record on which Armstrong’s scat singing, wordless syllables traded with the band, reached the public and, by the usual account, doubled as a commercial accident: the story that he dropped the lyric sheet and improvised the syllables is his own later telling, attached to the record because the record needed a legend, and it is reported here as his telling rather than as session fact. What is not in doubt is the effect. Heebie Jeebies sold, as the trade papers noticed, and from 1926 forward Armstrong the singer is a fact of the catalogue alongside Armstrong the cornetist, a doubleness the rest of this article follows to the end. The autumn sessions of 1926 show the group at its most various. Big Butter and Egg Man, from November, gives May Alix the comedy duet partner’s role and Armstrong a closing chorus that trumpet players studied for its construction, the melody’s own notes rebuilt into a staircase. Sunset Cafe Stomp, named for his nightly workplace, and Jazz Lips and Skid-Dat-De-Dat fill out a session that moves from dance tune to scat feature to ensemble romp without a weak side among them. Lil’s band, meanwhile, was recording at the Dreamland with her own groups, and the Chicago papers covered the Armstrongs as a musical power couple, two bandleaders in one household, an arrangement whose strains belong to the life article. For the catalogue, the point is simpler: by the end of 1926 the Hot Five had become a repertory company with a book of its own, repeat bookings at the studio, and an audience that bought the name. The workshop had become an institution, and institutions get to attempt bigger things, which is what 1927 brought.

One more feature of the first records deserves emphasis before the catalogue moves on: their humour is structural, not decoration. The spoken asides, the novelty instruments, the trombone’s commentary and the leader’s laughter are timed the way the solos are timed, and they teach the listener the group’s personalities the way a repertory theatre teaches its company’s. When the Hot Seven expansions and the 1928 sessions ask for graver attention, the listener grants it partly because the same voices that joked are now in earnest. Personality, in this catalogue, is not the enemy of art that purist criticism sometimes makes it; it is the medium through which the art is delivered, from 1925 to the last sessions.

How the Hot Five sessions were actually made

The Hot Five sessions were studio assemblies made quickly for a commercial series, which is the opposite of how canonical records are usually imagined. That is the third rung of the ladder, and it is worth pausing on the mechanics, because the myth of the Hot Five as a working band, honing its music nightly before an audience, is precisely backwards. The group did not play dances, did not tour and for most of its life did not exist between sessions. Its members had other jobs: Armstrong worked for Tate and then Dickerson, Dodds played around Chicago, Ory led his own groups and left for California for a stretch. Okeh summoned them when the catalogue needed titles, they recorded several sides in a day, and the band dissolved until the next summons. The records were made the way the blues accompaniments were made, fast, for money, inside the race series, and their casualness is part of their sound. Endings wobble, tempos sit where the day put them, and the joy in the playing is the joy of musicians amusing each other in working hours.

The commercial frame shaped the music’s content as well as its schedule. Okeh needed variety across a session: a blues, a novelty, a vocal feature, a stomp. Hence the spoken introductions, the comic exchanges on Big Butter and Egg Man, where the singer May Alix trades lines with Armstrong, and the instrument-switching display of Cornet Chop Suey, a showcase number whose title announces its purpose. Armstrong’s cornet feature records, Cornet Chop Suey above all, were advertisements for a technique, placed in the catalogue the way a piano roll of a virtuoso etude would be. Lil Hardin Armstrong’s compositions run through the sessions, Struttin’ with Some Barbecue chief among them, and her business sense runs through the paperwork: copyrights taken out, publishers chosen, the group’s name kept in front of the buyers. The Armstrongs’ marriage was ending by the last of these sessions, but the enterprise they had built together, a leader’s name on a label and a body of records under it, was the template for the rest of his career.

Technology changed under the sessions as they went. The first Hot Five dates were made as the industry converted from acoustic recording to the electrical process, with microphones replacing the recording horn, and the 1926 and 1927 sides gain a warmth and a low end, St. Cyr’s banjo and the tuba of the expanded group registering in a way the horn could not have caught. The three-minute limit of the ten-inch 78 governed everything: arrangements are compressions, introductions are short because choruses are precious, and the Armstrong solo architecture, statement and intensification inside two or three choruses, is partly an artefact of the format’s arithmetic. Understanding that arithmetic matters for the later argument of this article. When the format changed, the architecture stretched, into the nine-minute St. Louis Blues of 1954 and the two-minute single of 1964, but the shape of the thinking, exposition under pressure toward a climax, did not change with it.

Disputed entries cluster in these sessions too, and this is the place to name the pattern once. The vocalist on a given side, the banjoist’s identity on a December 1927 date where the guitarist Lonnie Johnson joined the group for Hotter Than That and Savoy Blues, and the exact division of piano duties on a couple of titles have all been argued by discographers, with the accepted discographies, and Riccardi’s corrections of them, resolving most and leaving a residue. The residue is reported here rather than resolved by fiat, in line with the rule this article set itself at the start. The takes themselves are the method’s best fossil record. Okeh generally recorded two takes of a title, issued one, and kept or discarded the other by the economics of the master vault; enough alternates survived the vault’s history to let later listeners compare, and the comparisons are decisive on a point this article cares about. Armstrong’s solos change between takes. The famous breaks and choruses are not fixed routines reproduced for the microphone but constructions rebuilt in the moment, sharing an architecture and differing in their detail, which is the signature of genuine improvisation operating inside a prepared design. The issued master was the take the session’s supervisor judged strongest or safest, and the alternates show the risk that the issued sides conceal: this music was being invented at speed, in a room, for money, by professionals who understood that the first duty of an invention is to land. That the landing rate was so high is the Hot Five’s real miracle, and it is a professional miracle, not a mystical one.

The studio’s economics also explain the catalogue’s early silences. Okeh recorded Armstrong when the Chicago operation’s schedule and the race catalogue’s needs aligned, which meant clusters of sessions separated by months in which the band’s development went unrecorded. The diary of the records is therefore a diary with pages missing, and the listener who plots the Hot Five’s growth from the sides alone is reading a sampled history, not a continuous one. The sampler’s gaps matter to the decline argument later, because the same sampling, applied to the big band years by inattentive reissue programmes, manufactured much of the evidence for a fall: judge any career by its worst-preserved fraction and a fall is what the archive will show.

The Hot Seven and the expanded group

In May 1927 Okeh expanded the studio group for a concentrated week of sessions, adding Pete Briggs on tuba and Baby Dodds on drums to the Hot Five’s five, and putting John Thomas in the trombone chair Kid Ory had vacated, and the records made under the name Hot Seven show what the added weight was for. The tuba and drums gave the group a rhythm section with a floor to it, and Armstrong’s playing changes to meet the new support: the solos lengthen, the phrases stride further across the bar lines, and the records begin to be built around featured choruses rather than shared ones. The week produced a run of titles that the later canon seized on entire. Willie the Weeper and Wild Man Blues open the run with the group’s new swagger. Alligator Crawl, a Fats Waller composition, rolls on Briggs’s tuba. Melancholy Blues slows the machinery to a procession. Keyhole Blues gives Armstrong a vocal of rueful comedy. And at the centre of the week sits Potato Head Blues, the record musicians cited for decades when they wanted to explain what he could do.

Potato Head Blues is built like the mature Armstrong solo in miniature. After an opening of ensemble and a clarinet chorus from Johnny Dodds, the band drops to stop-time, the rhythm instruments striking isolated chords and falling silent between them, and Armstrong takes two choruses alone against that scaffolding. The solo is a single rising argument: phrases that begin as variations on the melody climb into figures that leave it behind, the rhythm displacing and doubling, until the final chorus arrives at a peak that feels less like decoration than like a verdict. Nothing on record before it had given a soloist that much exposed space and heard him fill all of it. The record’s commercial fate is part of its meaning too: it was one side among many in the race catalogue, sold without fanfare, and its canonisation was the work of later listeners, musicians first and critics after them, who kept returning to those two choruses as a standard of what improvisation could be.

The expanded group lasted only for that season of sessions, and the reasons are instructive. The Hot Seven was not a band anyone could book; it was a recording configuration, assembled because Okeh wanted Armstrong records with a fuller sound as electrical recording spread and as his name began to carry a series. Its members dispersed back to their working bands when the dates ended. What persisted was the lesson of the configuration: that his playing wanted a rhythm section built to support a soloist, and that the soloist, given the support, would reorganise the music around himself. The big band arrangements of the 1930s and the All Stars format of the 1940s are both, in different ways, attempts to institutionalise what the Hot Seven week discovered by accident. The discovery also marks the point at which the records’ centre of gravity becomes undeniable. In the Oliver band the ear went to the ensemble and found Armstrong. In the Hot Seven the record is aimed at him from the first bar, and the ensemble has become the setting for a jewel that everyone in the studio, and everyone at the label, understood to be the product.

Personnel for the May 1927 dates is accepted in the main: Briggs, Baby Dodds and Thomas are the additions the discographies print. The trombone chair is the one place where older discographies wavered, some printing Ory out of habit because the Hot Five name implied him, and the correction to Thomas is a small, clean example of how the Armstrong discography has been steadily repaired by returning to the label files rather than to the legend. Two of the week’s records show the expansion working in opposite directions. The Wild Man Blues, taken at a tempo that turns its title ironic, is all space: long tones, holes in the texture where the tuba and drums breathe, a solo architecture of held notes and sudden runs that trusts the listener’s patience completely. Twelfth Street Rag, at the other extreme, compresses a piano rag’s strains into three minutes of ensemble brilliance, Dodds and Armstrong trading the lead in a display of collective virtuosity that looks back to the Oliver band and improves on its precision. Between the two, the week’s session book holds Willie the Weeper, Keyhole Blues, Melancholy Blues, the record later chosen for the Voyager Golden Record, and Gully Low Blues, Armstrong’s own slow blues composition, whose vocal returns to the accompanist’s art with himself as the singer accompanied. The Hot Seven week is usually summarised by its hit, Potato Head Blues, but the week’s real significance is its range: in five days of studio time the group had demonstrated ballad patience, rag precision, blues depth and comedy, a portfolio no other recording band of 1927 could have matched, and it did so with two borrowed chairs and no rehearsal tradition beyond the leader’s ear.

The week also settled the question of who the records were for. Okeh marketed the Hot Seven sides within the race series, but the musicians who bought them, white and Black, in Chicago and in the territory towns, were already treating Armstrong’s records as textbooks, and the trade papers noticed the crossover of influence before the crossover of sales. A textbook is a different commodity from a dance record: it is bought to be studied, replayed and argued over, and its maker’s authority compounds with each hearing. The authority Armstrong carried into 1928, and eventually into the general market, was built in those study sessions by other musicians, the one constituency whose verdict on technique cannot be spun, since their own playing had to answer for what they claimed to hear.

The records that changed the ensemble balance

Between the Hot Seven week of May 1927 and the end of 1928, a run of Armstrong sessions altered the basic contract of the jazz record: who leads, who accompanies, and where the listener’s attention is meant to live. The change can be heard across a sequence of dates. In December 1927 he recorded with the guitarist Lonnie Johnson as a guest, and the two sides from that encounter, I’m a Ding Dong Daddy in the company bookkeeping of the catalogue and, more lastingly, Hotter Than That and Savoy Blues, show the cornet in dialogue with a single stringed instrument, trading phrases and, on Hotter Than That, trading scat syllables with Johnson’s guitar figures in a passage that turns the record into a contest between friends. The same December brought Struttin’ with Some Barbecue, Lil Hardin Armstrong’s composition, on which her husband’s lead is so commanding, so rhythmically placed, that generations of trumpet players learned it as a text. The ensemble still plays the New Orleans parts, but the parts have become a frame, and everyone in the record seems to know it.

The working base underneath the change was the Sunset Cafe band of Carroll Dickerson, in which Armstrong played nightly in Chicago with a larger, reading ensemble, and when the Dickerson band began appearing on his records in 1928, under names that present him as the leader, the balance tipped for good. A Monday Date and Fireworks, from June 1928, are dance-band records reorganised around one man’s entrances. The arrangement provides riffs, modulations and stop-time breaks, and the cornet treats the arrangement as weather to fly in. This is the hinge of the early catalogue, and it is also where the story told about his whole career goes wrong if it is read as decline-in-embryo. The move from collective improvisation to featured soloist over an arranged band is the foundational move of jazz’s next forty years; Armstrong made it on record before anyone else, and the records of 1928 are the minutes of the meeting.

Two further records of that summer mark the shift in the rhythm of the music itself. Skip the Gutter and Muggles, from the June and December sessions with Earl Hines, loosen the four-beat foundation into something with air in it, Singleton’s drumming colouring rather than merely marking time. And the vocal records multiply: on Squeeze Me and on the December sides he sings entire choruses as a matter of course, no longer as novelty but as a second instrument the band expects. By the end of 1928 the Armstrong record has its mature grammar: an arranged frame, a lead voice that is doubled between horn and throat, and solo space allocated as the record’s reason for existing. The 1928 group’s sessions with Hines, to which the next three sections are given, are where that grammar produced its masterpieces. The December sides with Johnson deserve a closer description, because they show the dialogue method searching for new partners. On Hotter Than That, Johnson’s guitar answers the trumpet phrase by phrase in the breaks, and the record’s famous closing passage has Armstrong scatting against the guitar in an exchange so quick that the two instruments seem to be finishing each other’s thoughts. Savoy Blues gives the guitar the opening statement, a slow blues exposition that Armstrong then answers with a chorus of weighted simplicity, and the record’s patience, its refusal to hurry toward virtuosity, looks forward to the ballad records of the 1930s more than back to the Hot Five’s comedy. Johnson was the most modern guitarist in the music, the first to play single-note solos in the horn’s manner, and the sessions document the moment the horn’s manner and the guitar’s met. That the sides were issued as novelties of casting, the race catalogue’s habit, rather than as the chamber music they are, is another small exhibit for section three’s argument: the catalogue’s own packaging consistently undersold the art it contained.

The workplace behind these records matters as well. Armstrong’s evening job through these years was the Sunset Cafe on Chicago’s South Side, where he played for a floor show, the Plantation Revue, in a band led by Carroll Dickerson, while Lil’s group held the Dreamland across town. The records and the floor show fed each other: the showmanship, the announced features, the comedy timing that listeners hear in the Hot Five’s novelties were honed nightly before an audience that had paid to be entertained, and the Hot Seven’s ease, its sense of a performance staged for a room, comes from that stage. Catalogue histories that separate the records from the revue lose the context that explains their confidence. The Hot Five records were made by working entertainers at the top of a tough trade, and their joy is professional joy, which is why it survives the recording horn’s primitive ear.

The 1928 sessions with Earl Hines

The sessions Armstrong made for Okeh across 1928, in June, in the first days of July and again in December, with Earl Hines at the piano, are the summit of his early catalogue and the records against which the whole continuity argument of this article is first tested. Hines was the musical director of the Sunset Cafe band and a virtuoso of an entirely new kind: a pianist who played the keyboard the way a trumpet plays, with a right hand that stated melodies in octaves and a left hand that punctuated, answered and provoked. Put two such players in one studio band and the geometry of the music changes. The accepted personnel for the core sessions is Fred Robinson on trombone, Jimmy Strong on clarinet and tenor saxophone, Mancy Carr on banjo, Zutty Singleton on drums, Hines on piano, and Armstrong, now on trumpet rather than cornet, in front. The records they made in June include Fireworks, A Monday Date, Skip the Gutter and Sugar Foot Strut. The July dates added Squeeze Me and the record to which the next section is devoted. December brought Beau Koo Jack, Save It Pretty Mama, the Weather Bird duet, St. James Infirmary and Tight Like This, the last blues choruses of the series closing the year at a pitch of invention the race catalogue had never carried.

What distinguishes the sessions is reciprocity. In the Hot Five the supporting players are excellent and the record is still a soloist with accompaniment. With Hines present, every Armstrong chorus exists inside a conversation: the piano answers the trumpet’s phrases, anticipates its breaks, fills its silences with runs that amount to a second lead part, and twice, on the records examined in sections fourteen and fifteen, steps forward as an equal. Singleton’s drumming works the same way, accenting the soloist’s figures rather than merely keeping the floor level. The effect is a band that behaves like a chamber group of strong personalities, and it explains why these records, rather than the equally fine Hot Five sides, became the canonical Armstrong for musicians: they show the mature method entire, dialogue as structure, at the highest level of execution the recording technology of the day could capture.

Two honest qualifications belong here. First, the sessions were as commercial as everything else he made: Okeh sold them in the race series, some were issued under the name of the Savoy Ballroom Five or credited to the Dickerson band in its various billings, and their elevation above the rest of the catalogue is later work, done by reissue editors and critics. Second, the personnel record for December 1928 has its contested entries, as flagged in this article’s method: the identity of the band on St. James Infirmary, issued as Louis Armstrong and His Savoy Orchestra and drawn from the Dickerson orchestra, is secure in outline, while a handful of chair assignments on the December dates have been argued between the older discographies and the corrected ones, and the piano credit on individual June titles has had its own literature. None of the disputes touches the two principals. On every side that matters here, the trumpet is Armstrong and the piano is Hines, and the dialogue is unmistakable.

The 1928 band did not last. Hines left to lead his own orchestra at the Grand Terrace, Armstrong left Chicago in 1929 for New York and the theatre work that sections seventeen and eighteen describe, and the group became a might-have-been that musicians mourned for decades. The records remain as the proof of concept: for one season, the most advanced soloist and the most advanced pianist in jazz worked in the same band, and the catalogue caught it. The June 1928 dates also continued the accompanist’s work in a new key, when the same studio group, billed as his Hot Four, backed the singer Lillie Delk Christian, a cabaret vocalist whose sweet delivery Armstrong’s trumpet comments on with a wit that the blues singers’ gravity never invited. The pairing is worth a catalogue’s notice because it shows the accompanist’s art adapting to material at the opposite pole from Bessie Smith: where the Smith sides required weight, the Christian sides require sparkle, and the same horn supplies both by adjusting its commentary to the singer’s temperature. Parallel work of this kind, running alongside the headline sessions throughout the Chicago years, kept the dialogue method exercised in every register, which is part of why the method was ready, in December, to produce Weather Bird: the duet was not an experiment but a concentration of habits practised weekly for four years.

The autumn sessions of 1928 complete the year’s arc. Knee Drops, from July, is the new band at its most ebullient, a performance whose ensembles swing with a looseness the Hot Five never attempted. The September sessions produced sides with the singer and the small group in various configurations, and the year’s final Chicago dates, in December, gave the Weather Bird duet and the last of the Hines collaborations in the acoustic city’s studios before the New York move. Taken together the 1928 sessions form the most concentrated body of master recordings Armstrong ever made, a dozen titles that between them contain the cadenza, the duet, the dialogue and the ballad in embryo, and the year’s achievement should be stated without the mythology that usually attends it: these were working sessions, made for a race catalogue by professionals on dates, and their greatness is the greatness of craft operating at its highest pitch inside ordinary commerce. The commerce did not prevent the art. It never had.

West End Blues and its opening

Why is West End Blues considered important?

West End Blues, recorded on 28 June 1928, opens with an unaccompanied trumpet cadenza, some fifteen seconds in which Armstrong states the language his later solos would use: wide intervals, held notes pressed harder, phrases that turn back on themselves. The band then plays King Oliver’s blues as a slow ceremony, and Armstrong closes it singing wordlessly against the clarinet.

The composition was Oliver’s, and Armstrong had played the tune’s world from inside the Oliver band five years earlier, which gives the record its peculiar charge: the student performing the teacher’s blues as the new music’s manifesto. The structure is a twelve-bar blues taken at a funereal, stately pace. After the cadenza, the ensemble states the theme; Fred Robinson takes a trombone chorus of long tones; Hines plays a chorus of piano that decorates the harmony without disturbing its gravity; and Armstrong returns, first on trumpet and then, in the last chorus, as a singer without words, humming and scatting against Jimmy Strong’s clarinet in a duet that floats free of the beat while Singleton’s cymbal marks time beneath it. The record ends not with a climax of volume but with a rhapsodic piano figure from Hines and the band’s final chord, and the whole performance lasts barely three minutes. Its compression is the point. Every resource the 1928 group had developed is present, each one allotted exactly its space, and nothing is wasted.

The reception history divides cleanly, and the division is instructive for the argument of this article. On release, West End Blues was a respectable race-catalogue side, one of several Armstrong records that season, reviewed almost nowhere that print survives. Its canonisation began in the 1930s among musicians and the French critics, and was completed after the war by writers with access to reissues: Gunther Schuller’s analysis in Early Jazz, the most influential technical reading the record has received, treats the cadenza as a proclamation that reorganised what a jazz solo was permitted to be. That reading is deserved. But note what the reception history proves: the record’s stature was assembled retrospectively, by the same critical machinery that was simultaneously demoting the big band records without listening to them. Canonisation and dismissal were one process, and West End Blues was its beneficiary rather than its exception.

As a performance, the opening cadenza remains the crux. It is not an introduction in the ordinary sense, not a flourish to clear the throat, but a complete solo statement placed before the tune it belongs to, the argument announced before the evidence. Trumpet players studied it the way string players study a cadenza in a concerto, and its phrases entered the common vocabulary of the instrument. The claim this article makes for it is narrower than the legend and stronger: the cadenza is the 1925 method perfected, the paraphrase-and-intensify architecture of the accompaniments and the Hot Five solos, now executed with a command of time and tone that the earlier records promise and this one delivers in full. The supporting performances inside the record repay the same attention, because West End Blues is a group record crowned by an individual. Hines’s chorus, played in the stride-derived manner he called trumpet piano, answers the cadenza’s authority with an authority of another kind, figuration that imitates the horn’s articulation so closely that the two solos read as a debate between equals about how a melody may be decorated. Zutty Singleton’s drumming in the final chorus, brushes and cymbal colour behind the vocal, is the new band’s rhythm made audible, light, swinging and placed entirely at the service of the singer. And the ensemble’s closing statement, after Armstrong’s last vocal phrase, lands with the weight of a band that knows it has made something. Oliver had recorded his own composition that same season with his Dixie Syncopators, a respectable record that the Armstrong version instantly eclipsed, and the eclipse is the record’s final lesson: the composition was available to everyone, and what Armstrong added was not notes but conception.

The record’s institutional honours followed the critics’ verdict with the usual lag. West End Blues has been admitted to the Grammy Hall of Fame and placed by successive polls of musicians and historians at or near the head of any list of the music’s essential records, and its cadenza is taught, by ear, to trumpet students who may never learn another Armstrong solo. A catalogue article should note what such canonisation does to listening: the record becomes a monument, approached with reverence, and the comedy of its companions, the laughter in Muggles and Fireworks from the same sessions, falls out of the story. Heard in its session sequence, as the discographies arrange it, West End Blues recovers its context, one great record among the dozen its makers cut that summer, which is how its makers heard it and how it sounds most alive.

Weather Bird as a duet

On 5 December 1928, in the middle of the December sessions, Armstrong and Earl Hines recorded Weather Bird as a duet for trumpet and piano, and the record is the logical end of the year’s experiment: the band, having been reduced by stages from an ensemble to a frame to a conversation, is reduced at last to two men. The tune was Oliver’s Weather Bird Rag, which Armstrong had recorded with the Creole Jazz Band in April 1923, his first month on disc. Five years later he plays it as a dialogue with the one musician alive who could match him figure for figure. The performance is a chain of traded choruses and shared ones, Hines’s right hand taking the trumpet’s phrases and returning them altered, Armstrong answering with variations that quote the piano’s runs back at it, the two of them accelerating the old rag’s stride into something like chamber music played at dance tempo.

The duet clarifies what the Hines partnership actually was. Pianists before Hines accompanied; Hines converses, and Weather Bird is the conversation with the rest of the band sent out of the room. Its most famous passage comes near the end, where the exchanges tighten into short traded figures, each man capping the other’s phrase, a structure Armstrong would use for the rest of his career in dialogue with trombonists, singers and his own voice. Listeners who know the later records will hear the template at once: the exchanges with Jack Teagarden in the All Stars, the duets with Ella Fitzgerald in the 1950s, even the sung exchanges with Velma Middleton, all run on the circuitry this record installs. The through-line of the catalogue, which this article exists to trace, is rarely as audible as it is here, in a three-minute duet cut as a catalogue filler at the end of a long day’s session work.

The same December dates produced records that widen the picture of what the 1928 group could do: Beau Koo Jack, a flag-waver of controlled ferocity; Save It Pretty Mama, a slow blues performance; St. James Infirmary, issued under the Savoy Orchestra name, on which Armstrong sings the old lament over the Dickerson band’s arrangement with a theatrical gravity that points directly at the popular-song career of the 1930s. That a single week’s sessions could hold the duet, the flag-waver and the sung lament, all at the same level, is the strongest evidence the early catalogue offers for the continuity thesis: the range the later career is accused of dispersing into was present, entire, in 1928, waiting only for formats and markets that would pay for its parts separately. The rag itself carries a small history that the duet quietly revises. Armstrong had played Weather Bird in the Oliver band, where it functioned as an ensemble feature in the New Orleans manner, five voices trading the melody’s fragments. The duet strips away the ensemble and leaves the fragments in two hands, and the astonishing thing is how little is lost: the tune’s identity, which in 1923 lived in the collective texture, turns out to live in the phrases themselves, portable into any instrumentation the two men chose. This is the same portability the whole catalogue will demonstrate, the method surviving every change of band, format and market, and the duet is its first laboratory proof. The record also documents a working friendship at its peak. Armstrong and Hines were the two most advanced improvisers of their generation, Chicago’s twin attractions in 1928, and Weather Bird is the sound of each showing the other what he could do, competitively and affectionately at once. Their later reunion, when Hines joined the All Stars in 1948 and spent three years in the band, produced fine records, but the duet remains the purest document of the partnership, two musicians and a piano’s worth of orchestration between them.

As a piece of recording history the side has one more distinction: it was a commercial orphan that criticism adopted. Buried as the reverse of Muggles, a record whose title alone guaranteed it a certain fame, Weather Bird circulated quietly for years before the 1930s French critics and the LP reissue programmes elevated it into the inner canon. Its career repeats, in miniature, the catalogue’s larger pattern, in which the industry’s packaging and the music’s value travel separately, and in which the task of a serious catalogue is to keep reuniting them. The duet is also, not incidentally, the record that best answers the question of what to play a sceptic who knows Armstrong only as a gravel-voiced entertainer: four decades of argument, settled in under three minutes by two men at one piano.

What the 1928 group did that the 1925 group could not

The question opened in section nine can now be closed, and closing it lands the fourth rung of the ladder. What changed between the 1925 sessions and the 1928 ones is that the music moved from ensemble interplay to two soloists in dialogue, which is a different conception of the jazz record and not merely a better-executed version of the old one. The Hot Five of 1925 plays as a New Orleans band in good order: the cornet leads, the clarinet and trombone weave around the lead, the rhythm marks time, and the solo choruses, Armstrong’s included, are distributed as the form requires. The 1928 group plays as a hierarchy with a conversation at its top. The ensemble’s work is arranged to present two virtuosos, Armstrong and Hines, and the other instruments, however fine, function as a setting. Singleton’s drums comment. Robinson’s trombone provides a foil. Strong’s clarinet provides colour and, in the last chorus of West End Blues, a partner for the leader’s voice. Everything is subordinated to the exchange between the trumpet and the piano.

Four enabling changes made the new conception possible, and they are worth separating because later accounts often merge them into a mist of inevitable genius. First, personnel: Hines was simply a musician of a class the Hot Five did not contain, an improviser whose technique invited answer rather than accompaniment. Second, the instrument: Armstrong had moved from cornet to trumpet during these years, gaining brilliance and carrying power, and the 1928 records exploit the new edge. Third, the technology: electrical recording captured the piano’s full range and the drums’ colour, letting a dialogue at that density register on disc. Fourth, the working band: the Dickerson organisation at the Sunset Cafe gave the studio group a nightly discipline the pure studio assembly of 1925 lacked, so the 1928 sides combine the freshness of session work with the cohesion of a band that had played the arrangements in public. Remove any one of the four and the records change character. Together they amount to a new answer to the question of what a three-minute jazz performance is: not a miniature dance set, but a staged encounter between personalities, with the band as theatre.

The consequences reach far beyond Armstrong. Every record after 1928 that is built as a frame for featured soloists, which is to say the mainstream of jazz recording for the next forty years, descends from the settlement these sessions reached, and the settlement was reached as catalogue work, between dance jobs, for a segregated record series. How the technique itself is constructed, the time feel, the paraphrase method, the building of a solo as an argument, is the subject of the companion study of Louis Armstrong’s musical style, and this article deliberately leaves the mechanics there. What belongs to the catalogue is the shape of the event: in three years, across the Okeh sessions, the jazz record’s centre of gravity moved from the band to the man, and it never moved back. The rest of Armstrong’s career, with its big bands, its small groups, its duets and its pop singles, is played out inside the settlement of 1928, on stages and in formats the settlement made possible. Balance requires the losses to be stated as plainly as the gains, since this section’s claim is comparative, not absolute. The 1925 group’s collective polyphony, the interweaving front line of the New Orleans manner, recedes in the 1928 records to a frame for the soloists, and something real went with it: the Hot Five’s best ensembles have a communal joy that the later group’s polish does not reproduce, and the revival movement of the 1940s, which rebuilt bands in the 1925 image, was responding to a genuine musical value and not merely to nostalgia. Armstrong himself, in the All Stars years, restored the front-line interplay as a deliberate archaism that was also a homecoming, Teagarden and Bigard weaving around him as Ory and Dodds once had. The catalogue’s shape, read generously, is not a replacement of the collective by the individual but an expansion and contraction across a lifetime, the soloist’s art learned inside the collective, practised at full stretch in 1928, and folded back into a front line when the small band returned. The 1928 group did what the 1925 group could not: it made records in which one musician’s conception organises everything. What it could no longer do, nightly, was pretend that the conception was a democracy.

The 1928 settlement also fixed the pattern of his working life in a way the catalogue should record. From 1928 onward the band, whatever its size, was built around him: his name over the group, his features at its centre, his allocation of space its method. The big bands of the 1930s, for all their size, are the 1928 group enlarged, a star and his frame, and the All Stars are the 1928 group restored to chamber scale. Even the duet records and the songbook albums obey the settlement, Fitzgerald’s partnership being the honourable exception that proves it, the one format in which the frame held an equal. A catalogue organised by personnel will therefore always mislead if it treats the groups as the story. The groups are the variable scenery. The story is the conception that walks through all of them unchanged, and 1928 is where that becomes undeniable.

The turn to popular song

In 1929 Armstrong’s recording career changed markets, and the change is where this article’s longest open loop begins. How did a man in his sixties outsell the biggest band in the world? The question will not close until section twenty-eight, but its answer starts here, with the decision, his and his employers’, to aim the mature method at the general popular song. The move is usually described as the beginning of the decline, the moment the artist became an entertainer. The records suggest a colder calculation. The race catalogue was a ceiling as well as a home: its sales were capped by its distribution, its sessions paid flat fees, and its audience, however loyal, was the audience the industry had assigned him. The popular song market had no ceiling. A trumpeter who could sing, and whose singing by 1929 was as distinctive as his playing, could compete for the whole country’s ear, and the songs themselves, the new standards being written for Broadway and Hollywood, needed interpreters who could make a melody indelible in one hearing. That was precisely the skill the accompaniment years and the Hot Five had built.

The first records of the turn show the method transferring intact. I Can’t Give You Anything but Love, recorded in New York in July 1929 with Luis Russell’s orchestra behind him, treats a Broadway song the way the Hot Five treated a stomp: a vocal statement, a trumpet paraphrase that raises the stakes, a closing chorus that settles the account. The performance made the song a standard as much as the show it came from had done. In Los Angeles the following year, working with the Les Hite band at the Cotton Club and with a young Lionel Hampton on drums, he recorded Memories of You and You’re Lucky to Me, and the trumpet playing on Memories of You, a ballad performance of sustained, singing tone, announced that the ballad, jazz’s slowest and most exposed form, was now within his method’s reach. The ballad line that starts here runs unbroken to the records of his sixties.

The turn had a cost accounting that the decline narrative gets half right. The popular market wanted the singer and the personality at least as much as the trumpeter, and the sessions of the 1930s ration his horn more carefully than the Okeh sessions had. What the narrative misses is the direction of the transaction. Armstrong was not abandoning a serious art for money; he was selling the same art into a bigger room, at prices the race catalogue could not pay, and banking the difference in a fame that later let him record the Handy and Waller songbooks, the Fitzgerald duets and the concert albums on his own terms. The loop opened in this section stays open across the big band years and the All Stars, because its answer is the whole arc: the man who outsold the Beatles in 1964 was built by the turn of 1929, and the turn was a professional’s decision, made by the most employable musician in the business, with his eyes open. The Russell band deserves a sentence of its own in this account, because the 1929 Okeh sessions are joint work. Luis Russell’s orchestra was the strongest of the New York territory bands Armstrong recorded with, a band full of alumni of the New Orleans and Chicago scenes, Henry Allen among its trumpets, Albert Nicholas among its reeds, and an arranger’s sense of how to frame a soloist with riffing backgrounds that answered rather than covered him. Mahogany Hall Stomp, recorded with the band in March 1929, became its joint showpiece, a performance built around Armstrong’s breaks and the band’s answering figures, and it stayed in his concert programmes into the All Stars years, where it opens the Symphony Hall album like a signature. The Russell sides also show the catalogue’s division of labour in transition: the band provides the dance floor, the singer provides the song, and the trumpeter provides the event, three products in one three-minute package, each saleable to a different buyer. The Depression would shortly make such economy a necessity for every band in America. Armstrong’s records had practised it in prosperity.

The second thing the 1929 sessions show is the song machine’s hunger from the inside. Publishers placed songs with bandleaders the way salesmen place goods, and a singer who could make a lyric intelligible and a refrain unforgettable was courted with material of every quality. Armstrong’s recorded output from 1929 onward therefore reflects the machine’s output as much as his own taste, a fact the period criticism rarely allowed: the same sessions that produced Ain’t Misbehavin’ produced novelties nobody remembers, not because his judgement failed intermittently but because the system required volume. Sorting the wheat in the song years is the listener’s work, and the wheat is plentiful, which is why the turn to popular song cannot be treated as an artistic event at all. It was an industrial event. The artistry is in what the voice and the horn did once the machine delivered the material, and on that measure the song years are continuous with everything around them.

Ain’t Misbehavin’ and the theatre connection

The hinge record of the turn is Ain’t Misbehavin’, and its circumstances explain the theatre’s role in his catalogue. In the spring of 1929 Armstrong was in New York, and the revue Hot Chocolates, with a score by Fats Waller and Andy Razaf, opened at Connie’s Inn in Harlem with Armstrong in the orchestra pit, from where, by the account that has come down from the production, his trumpet interjections during Ain’t Misbehavin’ drew so much attention that he was brought forward to play the number in view of the audience. The revue transferred to Broadway, and on 19 July 1929 he recorded the song for Okeh with Luis Russell’s band, singing and playing it as a single continuous performance. The record is the template of the popular Armstrong: the voice enters as an intimate, slightly comic confidant, the trumpet answers as the master musician, and the two Armstrongs conduct the dialogue this article has been tracing since the Smith sessions, now within one man.

The theatre connection mattered beyond one song. Broadway and its Harlem feeders were the popular song industry’s research department, and a musician placed inside a revue heard the new material nightly, learned which songs held a room, and could carry them to the studios with an interpreter’s authority. Black and Blue, also from Hot Chocolates and recorded by Armstrong in the same July days of 1929, shows the reach of the education: a song about the interior of racial experience, delivered by Armstrong with a gravity that no amount of stage grin could be made to cancel, and received then and since as one of his defining vocal records. The theatre also taught scale. A revue audience is not a dance floor and not a cabaret table; it is a crowd to be held by personality projected across distance, and the projection, the timing of a laugh, the pause before a high note, became part of the recorded performances from 1929 onward. Critics who heard showmanship entering the records were hearing correctly. What they misjudged was the transaction’s price, because the playing inside the showmanship, on Ain’t Misbehavin’ itself, is as exact as anything in the Hot Five, the phrases placed with a jeweller’s economy.

Waller himself is part of the story this record starts. The composer of Ain’t Misbehavin’ and Honeysuckle Rose was Armstrong’s near contemporary and fellow professional, and the 1929 sessions began a relationship between interpreter and songbook that closes, twenty-six years later, with the Satch Plays Fats album of section twenty-five, on which the older Armstrong replays these very songs as classics. The catalogue folds back on itself in ways the two-career story cannot accommodate, and the fold begins in a Harlem theatre pit. The show behind the songs is worth its own paragraph, because the theatre connection was the catalogue’s bridge to the general market. Hot Chocolates opened at Connie’s Inn, a Harlem nightclub, in the spring of 1929, a revue with music principally by Waller and words by Andy Razaf, staged for the mixed downtown audience that Harlem nightlife drew. Armstrong joined the production as a pit musician and featured singer, and the producers, discovering what the pit contained, enlarged his part until the Ain’t Misbehavin’ feature became the show’s signature. The revue transferred to Broadway’s Hudson Theatre in June 1929, carrying the pit band and its singer with it, and the Okeh session of 19 July, with Luis Russell’s orchestra, put the theatre programme onto records while the transfer was fresh. The sequence matters: stage first, record second, the record as the souvenir and advertisement of a theatrical fact. From here to the end of the catalogue, Armstrong’s records repeatedly work this way, as the crystallisation of performances proven in rooms, which is one reason they wear so well. They were tested before they were sold.

The afterlife of the Black and Blue recording adds a literary witness to this article’s evidence. Ralph Ellison opened his 1952 novel Invisible Man with his narrator in a room full of light, listening to Armstrong’s record of the song and meditating on what the voice does with the lyric’s pain, and Ellison’s reading, the most famous passage ever written about an Armstrong record, treats the 1929 performance as a work of art entire, its irony and its suffering held in the grain of the singing. Ellison was not a jazz critic conserving a canon; he was a novelist identifying a master. The witness is worth citing in a debate that sometimes proceeds as if only specialists defended the big band singer: the century’s great American novel of Black interior life chose the supposed entertainer’s record as its emblem of art made inside invisibility. The catalogue’s standing outside the jazz press, in the culture at large, was set early, and it was set by the records the period division underrates.

The big band recordings of the 1930s

From 1930 until the middle of the following decade Armstrong recorded almost entirely in front of big bands, and this is the stretch of the catalogue the canonical division teaches listeners to skip. The records divide into groups by employer and city. There are the Los Angeles sessions of 1930 and 1931 with Les Hite’s Cotton Club orchestra, where Hampton’s drums drive I’m Confessin’ and If I Lose Let Me Lose. There are the Chicago sessions of 1931 and 1932, with a band organised under Zilner Randolph’s musical direction for Victor, producing When It’s Sleepy Time Down South, which became his theme, and Shine, a bravura performance that packs a comic vocal, a trumpet display and an accelerating final chorus into three minutes. There are the Victor sessions of 1932 and 1933, I’ve Got the World on a String and Hustlin’ and Bustlin’ for Baby among them, on which the band is anonymous and polished and the leader treats current pop songs the way he had treated the Henderson arrangements, as frames built to be broken into by the soloist. And from 1935, under the management of Joe Glaser and a Decca contract, there is the settled big band of the later thirties, Luis Russell’s orchestra in all but name, with Russell as musical director and Armstrong out front as singer, trumpeter and master of ceremonies on record.

The Decca records repay the listening the canon withheld. I’m in the Mood for Love and You Are My Lucky Star in 1935 show the ballad method at big band scale, the voice confiding over a cushion of reeds and the trumpet entering like a second, surer self. Jubilee, recorded in 1938 from the film Every Day’s a Holiday, swings with a looseness the Victor band never managed. Jeepers Creepers, from the film Going Places, made a hit of a Harry Warren song by sheer rhythmic address. And on 13 May 1938 the band recorded When the Saints Go Marching In, the performance that carried a sacred standard from the church and the funeral parade into the swing repertoire, where it has remained as his signature ever since. The Saints was not a novelty to him; it was New Orleans funeral music he had grown up inside, and the record’s authority comes from that source. Critics of the day treated the big band records as show business, and some of the product was: the Decca schedule demanded volume, and not every session’s material deserved him. But the trumpet choruses embedded in the ordinary dates, the breaks and codas he was handed and transfigured, hold up under the same scrutiny as the Okeh work, at higher fidelity and against bigger competition.

Two structural facts explain the sound of the decade. The first is rationing: a dance band arrangement gives its star two choruses where a Hot Five side gave him the record, and the art of the 1930s Armstrong is an art of compression, of making eight bars carry what thirty-two had carried in 1928. The second is the voice’s promotion. By the late 1930s he sings on most records and plays on all of them, and the singing has developed into an instrument with its own technique, the gravel register, the rhythmic displacements, the habit of paraphrasing a lyric’s rhythm while keeping its sense, which singers in every later popular genre studied in him whether they knew the source or not. The decade’s strangest single record may be Laughin’ Louie, cut for Victor in April 1933, a novelty in which Armstrong’s laughter overtakes the performance until the band itself dissolves, issued while his affairs were at their most chaotic and his management in dispute. Critics of the period filed it as clowning, and the filing is understandable; but the record belongs in a catalogue for what it documents about the instrument. The voice at play, timing laughter against a beat, registering changes of pitch and attack inside the joke, is the ballad method’s machinery running without a song to carry, and the machinery is exquisite. The Victor period closed with the band’s records growing stronger as the arrangements settled, and the Armstrong who sailed for England in 1932 and toured through 1933 and 1934 was, by the testimony of the musicians who heard him nightly in those years, playing with a power the records of the period only intermittently capture. The gap between the nightly evidence and the recorded evidence is the decade’s standing warning against catalogue judgements made from records alone, and it applies with the same force in both directions: the records are the evidence we have, and they are not all the evidence there was.

What the big band format gave him

It is easy to list what the big band cost: solo space, small-group interplay, the freedom of the studio assembly. The format’s gifts are less often itemised, and the catalogue is unintelligible without them. The first gift was dynamics at scale. A dance orchestra of fourteen pieces could be arranged in terraces, reeds against brass, full band against rhythm, and Armstrong’s entrances could be staged by orchestrators the way an opera stages an aria: the Randolph and Russell arrangements learn to hold the band back, let the voice confide, and then release the full ensemble under the trumpet’s climactic chorus, a dramaturgy the Hot Five’s democracy could not produce. The second gift was repertory. A working big band lived on the current song output of Broadway and Hollywood, and fronting one put Armstrong inside the standards at the moment of their manufacture. The songs he recorded in the 1930s, the Gershwin, Porter, Berlin and Warren titles that fill the Decca ledgers, became the shared language of American music, and his versions, however the critics ranked them at the time, were among the versions by which the country learned the songs.

The third gift was professional survival through the industry’s worst years. The Depression collapsed the race catalogue economics that had built his early career; labels folded, series were cancelled, and musicians of his generation were stranded by the hundreds. The big band format, with its ballroom circuit, its film shorts and its radio remotes, kept Armstrong employed, visible and recording through the whole decade, and the Glaser management, whatever its later critics have said about its commercial priorities, kept the band on salary and the records coming. A catalogue article should record that plainly: the 1930s records exist in quantity because the format was an economic engine, and the later career, the All Stars, the songbooks, the duets, was built with capital, financial and reputational, that the big band years accumulated. The fourth gift was the audience itself. The Decca records played in homes that had never bought a race catalogue side, on jukeboxes in towns without a Black record shop, and the man who would fill concert halls in the 1950s and top the chart in 1964 was introduced to that public here, singing I’m in the Mood for Love over a dance band, not leading the Hot Five.

None of this requires pretending the decade’s records equal the 1928 sessions as a group. They do not; no decade of any musician’s records does, and the argument of this article has never needed them to. It needs only the narrower claim, testable on any of the dates named in the previous section: that the technique is continuous, that the trumpet choruses of the big band years are the work of the same musician operating the same method inside a different rationing system, and that the rationing, not the man, is what changed. His own remakes supply the controlled comparison that clinches the section’s argument. In 1938, for Decca, Armstrong recorded Struttin’ with Some Barbecue again, the tune he had made famous with the Hot Five in 1927, now with the big band in a fuller arrangement and his own trumpet carrying a longer version of the argument the small group had made. Set the two performances side by side and the decline vocabulary fails again: the later solo is more spacious, its climaxes placed with a veteran’s patience, its rhythmic relation to the beat more teasing, while the band behind him supplies a weight the Hot Five could not. What changed between the versions is the frame and the allocation, the small group’s density traded for the big band’s power, and what persisted is the soloist’s architecture entire. Remakes are a catalogue’s gift to its historians, because they hold the material constant across the very years the critics divided, and Armstrong’s catalogue is full of them, the Saints, Ain’t Misbehavin’, Black and Blue, St. Louis Blues all recorded again and again. Each pair testifies the same way.

The Decca years also preserved the full range of his public personae in a way the race catalogue never could. The 1938 session that produced the Saints remake also produced a pair of recorded mock sermons, the Elder Eatmore performances, in which the preacher-comedian of his stage act addresses his congregation on generosity and diet with a timing that vaudeville would have envied, Armstrong the entertainer preserved on the same label, in the same season, as Armstrong the balladeer and Armstrong the trumpeter. The period criticism treated the personae as rivals, the entertainer diminishing the artist. The catalogue treats them as one instrument’s registers, and a listener who works through the Decca boxes in order hears what the audience of the time heard: a complete performer whose comedy, balladry and trumpet were one continuous act of communication.

The European sessions

Armstrong’s first European tour, in the summer of 1932, took him to London, where he played the Palladium and recorded a handful of sides with a pickup orchestra of British and continental musicians for the Parlophone label, and the trip’s significance for the catalogue is double. The records themselves, among them versions of Chinatown My Chinatown and You Rascal You, are capable big band performances in the period manner, valuable chiefly as evidence of how completely his method could animate musicians who had learned jazz from records, in this case from his. The reception was the revelation. European audiences and the young French and British critics treated the visitor as a major artist rather than as a race catalogue entertainer, and the Hot Club de France, founded in Paris in that same year of 1932, began the critical literature on his records that would, in time, set the terms for the American discussion too. The canon that later ranked his periods was, in embryo, a European invention, and it was invented by men listening to the whole man in front of them.

He returned in 1933 and 1934 for longer tours, and the Paris session of October 1934, made with an orchestra of American expatriates and French professionals for French Brunswick, produced the most substantial European records of the decade: a St. Louis Blues that revisits the Bessie Smith accompaniment from the leader’s chair, a Tiger Rag of controlled hysteria issued in France as Super Tiger Rag, and versions of On the Sunny Side of the Street and I Can’t Give You Anything but Love that test the popular method against the most sophisticated rhythm section he had worked with since Hines. The Paris sides have a documented looseness, the sound of a band enjoying itself at the edge of the arrangement, and they stand in the catalogue as the bridge between the Victor years and the Decca settlement: proof, made abroad, that the big band Armstrong could still produce records of the first rank when the material and the room were right. The European sojourn also ended in professional wreckage, a collapsed management, a lip injury aggravated by overwork and a period of rest and reorganisation in Chicago, out of which came the Glaser contract and the Decca years. The records remember the triumph; the biography, told in the companion life article in this cluster, remembers the cost. The Copenhagen visit of October 1933 left the period’s most valuable visual evidence: a film of Armstrong performing with his band before a Danish audience, among the earliest extended footage of him playing, in which the physical facts of his art are preserved, the stance over the band, the handkerchief, the announcements, and choruses taken at a heat that the touring schedule makes remarkable. The film matters to a catalogue article because it corrects the inference that the early 1930s records’ unevenness meant an artist coasting. In Copenhagen he is working at full stretch for an audience that had crossed a continent of expectation to see him, and the Swedish and Danish press treated the tour as a major cultural event. The Scandinavian reception also began the pattern that would repeat in 1949 and through the 1950s: Europe as the territory where the whole Armstrong, the legend and the working bandleader together, was received without the American criticism’s period anxiety. He noticed, and the loyalty ran in both directions for the rest of his career.

The working conditions explain some of what the records show. The British variety circuit demanded two houses a night in theatres where the band was one act among many, comedians, dancers, animal turns, and the Paris engagement that followed the London season was the first extended residency of his European years, in a city whose jazz public, organised around the Hot Club de France, had been studying his records for a decade. The French sessions of October 1934 were made for that public, and their repertoire, standards and earlier successes redone for French Brunswick, shows the repertoire consolidation that exile imposed: separated from the New York song machine, the book froze at the point of departure, and the records preserve the 1932 Armstrong in amber. Exile ended in January 1935, and the Decca contract began in the autumn, which is why this article counts the European years as the catalogue’s hinge: everything before them belongs to the first American career, everything after to the second, and the hinge itself is made of work, not of decline.

The All Stars recordings from 1947

In 1947 the economics and the aesthetics of Armstrong’s career changed together, and the catalogue’s second half begins in earnest. The big band business was collapsing under wartime costs and changing taste, Glaser’s office read the market, and Armstrong, forty-six years old and the most famous jazz musician in the world, was re-presented to the public as the leader of a small group. The trigger events were two New York concerts: a Carnegie Hall date in February 1947, whose second half, played with a small group rather than the big band, produced an audience response out of all proportion to its place on the programme, and the Town Hall concert of 17 May 1947, where a hand-picked small band behind him played the old repertoire as chamber music and the evening was recorded. On 13 August 1947 the All Stars opened at Billy Berg’s club in Hollywood: Armstrong, Jack Teagarden on trombone and vocals, Barney Bigard on clarinet, Dick Cary on piano, Arvell Shaw on bass, Big Sid Catlett on drums, with Velma Middleton, who had sung with the big band since 1942, as the featured vocalist. The name was marketing and the personnel was an argument: the front line of the New Orleans tradition, three horns and a rhythm section, applied to the whole of his songbook.

The early All Stars records, made for Decca from 1947, show the format finding itself. Rockin’ Chair, the Hoagy Carmichael song Armstrong and Teagarden sang as a duet of two old friends, became the group’s emblem, a performance whose easy affection is as constructed as anything he ever did and just as durable. Jack Armstrong Blues, credited jointly to Armstrong and Teagarden, takes the band through its paces as a calling card. The clarinet chair would pass to Edmond Hall in the middle 1950s, the piano chair to Billy Kyle for the long central period from 1953, the trombone chair to Trummy Young from 1952, and the drum chair through a line of players including Cozy Cole and Barrett Deems, but the design never varied: front line, rhythm, two singers counting Armstrong himself, and a book that mixed the New Orleans standards with the popular songs he had been gathering since 1929.

What the All Stars records do that the big band records could not is restore the dialogue. Teagarden was the one trombonist of the era whose lyricism matched the leader’s, and their exchanges, on record after record, run on the circuitry of the 1928 Hines sessions: statement, answer, escalation, reconciliation. Bigard’s clarinet supplied the third voice of the old front line, liquid and precise. For listeners who had filed Armstrong under entertainment, the All Stars’ first Decca albums were a recalibration, and the critics of the revival era, who had spent the 1940s demanding exactly this format from him, received the group as a vindication. The irony deserves noting in a catalogue that has tracked the canon’s construction: the All Stars were celebrated as a return to his beginnings by the same critical culture that had used those beginnings to demote his middle years, and the records themselves are better than either story, being not a return but a consolidation, the whole career, blues, standards, novelties and all, carried by six men. The All Stars’ middle period added records that fill the catalogue’s map of his genres. The 1951 Decca duet of Baby, It’s Cold Outside, with Velma Middleton, turned the group’s stage comedy into a hit record without a trumpet chorus worth the name, another exhibit for the voice’s independent career. Back o’ Town Blues and the group’s remakes of the New Orleans standards maintained the repertoire’s roots in public view. The band’s Decca album Satchmo at Pasadena, recorded in concert in 1951, added a second major live document to Symphony Hall, looser and more raucous, and the difference between the two concerts, the formal hall and the California auditorium, maps the group’s range of address: the same book, pitched to a black-tie audience as art and to a Saturday crowd as a party, both pitches honest. The group’s studio singles of the early 1950s, aimed at jukeboxes, are the catalogue’s least distinguished stratum, and candour requires saying so; the format rationed him hardest exactly where the money was quickest, and the records show it. The exception proves the rule that governs this whole article: the failures cluster where the format gave the method no room, and vanish where it did.

Personnel stability underwrote the group’s consistency. Young held the trombone chair from 1952 until his death in 1963, Kyle the piano chair from 1953, Hall the clarinet chair from the middle 1950s, and the rhythm section’s changes, Catlett to Deems in the drum chair, were absorbed by a book that every member knew by heart. A band that stable becomes an instrument the leader can play the way he played the trumpet, by instinct and allocation, and the All Stars records’ famous ease is the sound of that knowledge. Stability also made the group a conservatory: Young, Hall and Kyle were masters whose own styles were fully formed before they joined, and the band’s art was the diplomacy among formed styles, conducted nightly by the leader’s allocation of space. The catalogue’s late peaks, the Handy album above all, are unthinkable without that diplomacy.

The live albums

Armstrong’s catalogue contains a second, shadow discography of concert recordings, and from 1947 the shadow grows until it rivals the studio work. The foundational document is the Town Hall concert of May 1947, recorded by RCA Victor and issued as an album, on which the small group that preceded the All Stars plays with a relaxation and an audience rapport that the studio dates of the period do not match. In November of the same year Decca recorded the All Stars at Boston’s Symphony Hall, and the resulting album, Satchmo at Symphony Hall, drawn from the concert of 30 November 1947, became one of the best-loved live records in jazz: the band stretching the repertoire across full concert sets, the leader pacing an evening rather than a three-minute side, Muskrat Ramble and the standards unfolding at a length the 78 era never permitted. The difference the long form makes is the point. In the studio, the method was compression; on the concert stage, recorded whole, it becomes architecture in real time, choruses building across ten minutes with the crowd as a structural element, the applause and the laughter written into the performance’s rhythm.

The 1950s extended the live catalogue as touring became the centre of his working life. The Pasadena concerts, recorded at the Civic Auditorium in 1951 and again in the middle of the decade and issued by Decca, show the All Stars as a road-honed entertainment machine whose music loses nothing to its polish. The concert at Lewisohn Stadium in 1956 with Leonard Bernstein conducting and Mahalia Jackson appearing, filmed and issued as Satchmo the Great, placed him in the frame of American cultural prestige that the 1930s critics had reserved for his youth. The live records also preserve what the studio increasingly edited out: the full extent of the show, the comedy with Velma Middleton, the sermons and the set pieces, the pacing of a man who had been holding rooms since the Sunset Cafe. A listener who knows only the studio albums knows the musician. The live albums are where the entertainer and the musician are documented as the same person, which is the fact the two-career story exists to deny.

Judged as records, the best of the live albums also answer the decline narrative on its own preferred ground, the trumpet playing. Freed from the three-minute ration and the arranger’s frame, the Armstrong of the Symphony Hall and Pasadena sets takes solo after solo at concert length, and the playing has the range, the command of register and the rhythmic authority of the late 1920s work, applied now to material that includes the standards of his middle years. The live catalogue is the control experiment of this article’s argument: same man, same method, longer forms, and no cliff anywhere in it. The Symphony Hall album’s programme shows the concert idea at full stretch. The performance opens with Mahogany Hall Stomp, the 1929 showpiece renewed as a signature, proceeds through the repertoire’s whole range, Muskrat Ramble and Royal Garden Blues for the New Orleans inheritance, Black and Blue for the theatre years, ballads and blues between, and closes in the exuberance of a band playing for a hall that had come to celebrate as much as to listen. What the two LPs preserve, and what the studio records of the same years cannot, is proportion: the trumpet’s share of a working evening, the patience of the pacing, the way a feature for each sideman sits inside the leader’s architecture. The album became the All Stars’ defining document, the record against which the group’s studio work was measured, and its reissues have kept it in print in every format since, a continuity of availability that the 1930s records never enjoyed and that goes far to explain the late period’s critical standing: what stays in print gets judged, and what gets judged by performance rather than by reputation tends to vindicate this band.

The stage held one more advantage for the catalogue, and it appeared as the 1950s opened: tape. The concert recordings of the LP era could be made anywhere, and they were, in Pasadena and Chicago, at Newport in 1956 in the performance preserved in the film Satchmo the Great, and across the European halls the band worked each year. The accumulated live catalogue is uneven, as live catalogues are, but its best documents share the Symphony Hall quality of occasion, and its existence changes the evidence base for the late period: the critic of 1935 had the 78s and his memory of the dance hall; the historian of the All Stars has the evenings themselves, unedited, against which the studio rationing can be measured and understood as rationing. The playing, unmeasured and at full length, is the strongest single rebuttal the catalogue contains to the idea of a technique in retreat.

The Plays W C Handy project

On 12 July 1954, in a single day of sessions in New York, Armstrong and the All Stars recorded the album Louis Armstrong Plays W. C. Handy for Columbia, produced by George Avakian, and the record stands as the fullest answer the later catalogue makes to the charge of decline. The material was the songbook of William Christopher Handy, the composer of St. Louis Blues, Memphis Blues and Beale Street Blues, the published blues that Armstrong had accompanied behind Bessie Smith in 1925 when the idiom was new to records. The personnel was the settled All Stars of the period: Trummy Young on trombone, Barney Bigard on clarinet, Billy Kyle on piano, Arvell Shaw on bass, Barrett Deems on drums, and Velma Middleton taking the vocals on several titles. Avakian’s production decision was to let the performances run at LP length, and the album’s centrepiece, a St. Louis Blues of more than nine minutes, builds its trumpet choruses one on another until the performance has the shape of a concerto movement, the arrangement’s returns giving the soloist a structure to climb.

The album’s achievement is the union of the catalogue’s two supposed halves in a single session. The St. Louis Blues performance is in direct dialogue with the 1925 accompaniment: the same song, the same accompanist’s intelligence, now occupying the whole record, with twenty-nine years of technique between the two statements and no audible seam where invention is supposed to have stopped. Yellow Dog Blues, Loveless Love, Ole Miss, Chantez les Bas and the valedictory Long Gone fill out a programme that treats Handy’s blues as concert repertoire. Handy himself, old and blind, was played the record, and Avakian’s account of the composer’s response, delight mixed with the recognition that the music had outgrown the page he wrote for it, has become part of the album’s documented history. Critics who had spent decades rationing their respect received the Handy album with something close to surrender: here was a record by the fifty-two-year-old entertainer that the canonical criteria themselves, invention, command, structural ambition, rated at the top of his work.

For the catalogue’s shape, the Handy album also marks the moment the LP concept record enters Armstrong’s work as an artistic instrument rather than a repackaging. A single composer, a single band, a single day’s sessions, sequenced as a programme: the format had been pioneered in his catalogue only in the sense that his whole career had been a set of variations on songbooks, Handy’s in the blues years, Waller’s in the theatre years. Avakian made the variation explicit and gave it the long form it needed. The Handy album was a commercial success as well as a critical one, and its success commissioned its successor, the Waller album of the next section, and confirmed the template of the late catalogue: the songbook album, the duet album, the concert album, each a format built around the same undiminished method. A track-by-track pass shows the album’s architecture. St. Louis Blues opens the programme at full length, the statement of intent. Beale Street Blues gives Trummy Young a trombone feature that Young, the band’s powerhouse, turns into the album’s second summit, proof that the songbook format let the sidemen expand as it let the leader. Atlanta Blues, Handy’s version of the Make Me a Pallet song, features Velma Middleton in the blues-singer role this article traced to 1924, the accompanist’s art now practised by Armstrong behind his own band’s singer. Ole Miss reaches back to Handy’s earliest published blues, and Long Gone, the closing performance, sends the album out at a romp, the leader’s trumpet over the band in full cry. The sequencing is an argument in itself: blues as history, blues as feature, blues as duet, blues as farewell party, the songbook idea proving that one composer’s book contains a whole evening’s emotional range when the performers are equal to it.

The album also repaired a specific injustice of the catalogue’s middle years, and the participants knew it. Through the Decca period Armstrong had recorded Handy’s tunes as items in a variety programme, three minutes at a time, arranged for the market of the moment. The Columbia album returned to them with the space and the production seriousness of the LP age, and the reviews, Morgenstern’s among them, recognised the event for what it was: the music’s founding soloist and its founding published composer, joined at full length at last. Handy lived to hear it, and the photograph of the elderly composer listening, which the label circulated, became one of the era’s emblematic images of jazz honouring its own ancestry. Ancestry, though, is the wrong final word for an album this alive. The Handy set did not embalm the blues; it demonstrated that the blues book, in these hands, could still generate the longest and most searching performances of Armstrong’s recorded career, twenty-nine years after he first accompanied them.

The Satch Plays Fats project

Columbia and Avakian followed the Handy album in the spring of 1955, across two sessions, with Satch Plays Fats, a programme of Fats Waller compositions, and the pairing of the two albums states the late catalogue’s method as clearly as any manifesto could. Waller was the composer of Armstrong’s theatre breakthrough, and the album revisits the association at full LP scale: Honeysuckle Rose, Blue Turning Grey Over You, I’m Crazy ‘Bout My Baby, Squeeze Me, Keepin’ Out of Mischief Now, All That Meat and No Potatoes, I’ve Got a Feeling I’m Falling, and remakes of the two songs that anchor the whole story, Black and Blue and Ain’t Misbehavin’. The remakes are the album’s argument in concentrated form. Set the 1955 Ain’t Misbehavin’ beside the 1929 original and the comparison the decline narrative fears is the one the record invites: the tempo is more relaxed, the voice grainier and more knowing, the trumpet choruses built with a patience the young man, for all his fire, did not practise, the climaxes placed with the confidence of a musician who no longer needs to prove the instrument can reach them. Nothing is repeated and nothing is diminished. The performance is a reading, the second thoughts of the artist who made the song famous, and the catalogue is richer for containing both.

The band on the Waller sessions is the same All Stars personnel as the Handy date, with Velma Middleton again sharing the vocals, and the album shares its predecessor’s virtues: room for the front line to converse, arrangements that frame rather than confine, and a producer content to let a performance find its own length. Between the two songbook albums, a listener possesses the strongest single evidence for this article’s thesis, because the songbooks are the control condition of the whole inquiry. The material is held constant, compositions he had known since the 1920s, and the variable is time alone. If technique declined across those decades, the albums would show it in the plainest A and B comparison discography can offer. They show the opposite, and the critics who reviewed them at the time, Morgenstern prominent among them, said so in print, in the jazz press that had spent years qualifying its Armstrong coverage with the period division. The qualifications did not vanish, but after 1955 they had to argue with two Columbia albums, and arguments with records tend to be settled by the records. The Waller repertoire also carried a private charge that the album’s ease conceals. Waller had died in 1943, and the two men had overlapped for twenty years in the same Harlem world without ever making the records together that the songbook now imagines. Armstrong’s own memory of Waller, told in interviews across the years, was of the organist at the rent parties and the composer whose tunes he had plugged into theatre history in 1929. Satch Plays Fats is therefore a memorial as well as a programme: the surviving partner of an unwritten collaboration performing the dead man’s book with full honours, the All Stars as the orchestra of the occasion. That gravity is audible in the album’s pacing. The comedy is present, Waller is unplayable without it, but the tempos are considered, the trumpet statements full-length, as if the album understood itself as the definitive edition of a partnership that existed everywhere except on record.

Set beside the Handy album, the Waller set also clarifies the songbook idea itself, which Armstrong and Avakian effectively invented for jazz in those two years. The classical world had its complete editions and the pop world its composer tributes on radio, but the jazz LP of 1954 was young, and the notion that a single composer’s book, played entire by a single artist whose history was entangled with it, could constitute an album’s whole argument was new. The format that resulted, the songbook album, became one of the LP era’s standard genres, practised by Fitzgerald herself in her composer series for the same Verve stable a few years later. Armstrong’s two examples came first, and they came from the supposed decline, a fact the decline narrative has never digested.

The Ella Fitzgerald collaborations

Between 1956 and 1957 Armstrong recorded three albums with Ella Fitzgerald for Norman Granz’s Verve label, and they complete the late catalogue’s demonstration. Ella and Louis, made in Hollywood in August 1956 with the Oscar Peterson quartet, Peterson at the piano, Herb Ellis on guitar, Ray Brown on bass and Buddy Rich on drums, strips the duet format to its essentials: two voices, his trumpet, and the most unobtrusive rhythm section in jazz. The programme is the standards songbook, Can’t We Be Friends, Isn’t This a Lovely Day, Moonlight in Vermont, Cheek to Cheek, April in Paris, and the performances are duets in the sense this article has used since the Bessie Smith section: statement and answer, the accompanist’s art now practised by both parties on each other. Fitzgerald sings with a purity of intonation that exposes any roughness beside her, and Armstrong’s singing, placed against it, reveals its method nakedly: the syllables bent to the rhythm, the melody paraphrased the way his trumpet paraphrases, the gravel of the voice functioning as a timbre chosen from a palette rather than a defect endured. His trumpet choruses on the album are models of the ballad method that began with Memories of You in 1930.

Ella and Louis Again followed in 1957, a double album with Louis Bellson in the drum chair, extending the same conversation, and the third collaboration, Porgy and Bess, recorded the same year with Russell Garcia’s orchestra, took on the Gershwin score entire, Armstrong singing and playing across the roles, Fitzgerald at her most dramatic, Summertime and I Got Plenty o’ Nuttin’ among the performances that carried the album into the permanent catalogue. Fitzgerald’s own career, the songbook albums she was making concurrently for Granz and the stature that made her the one singer who could meet him as an equal in a duet, is told in the companion life of Ella Fitzgerald; here the point is what the sessions prove about him. These albums land the fifth rung of the article’s ladder: the Ella Fitzgerald sets and the W. C. Handy project are late works of the first rank, which the decline narrative cannot accommodate. A framework that must classify Ella and Louis as entertainment while classifying the Hot Five as art has classified the formats and the decades, not the performances. The duet albums stand in the catalogue where the Weather Bird duet stood in 1928, as the reduction of his art to its essence, two musicians in conversation, with the difference that the conversation now spans the whole of American song.

The Fitzgerald sessions also record a professional fact worth stating in a catalogue: the duet partner was the most technically accomplished singer alive, the sessions were produced by the most powerful impresario in jazz, and neither the partner nor the producer was in the business of charity toward a faded name. The late catalogue was made at the top of the industry’s market for quality, by people who could measure exactly what they were buying. Heard closely, the first album is a study in accommodation between two opposite vocal perfections. Fitzgerald’s instrument is all finish: intonation so exact that engineers used her takes to test equipment, a tone without grain, diction that sets every consonant in place. Armstrong’s is all character: the grain, the paraphrase, the syllable bent until it swings. The records work because each singer treats the other’s method as information. Fitzgerald, the most rhythmically exact of singers, relaxes her placement toward his; Armstrong, hearing that purity beside him, sings with a tenderness and a textual fidelity that his own albums do not always ask of him, and then takes a trumpet chorus that restores the balance by reminding the room where swing comes from. The opening track, Can’t We Be Friends, sets the pattern for the set, and by April in Paris and Cheek to Cheek the partnership has the ease of a long marriage. Critics of the day, including some who had filed Armstrong under entertainment for decades, reviewed the album as a summit, and its sales carried it into the general record-buying public that Fitzgerald’s songbooks and Armstrong’s singles had separately prepared.

The Porgy and Bess album extends the argument from duet to drama. Russell Garcia’s orchestral settings give the score its concert weight, and Armstrong’s assumption of multiple roles, the lover, the cripple, the villain’s tempter, tests the actor in him at full length. Bess, Summertime and the duet passages place the two voices inside Gershwin’s architecture rather than in front of a rhythm section, and the record holds its place in both artists’ catalogues and in the opera’s own performance history. A catalogue judgement that rates 1928 above all else must find room for the fact that the Armstrong of 1957 could be handed an opera and carry his half of it opposite the finest singer of the age, on a major label, for a producer who staked prestige money on the result. The late catalogue’s peaks are not survivals. They are conquests of new territory by the old method, which is what makes the decline story’s survival, in the face of them, a phenomenon of criticism rather than of music.

Mack the Knife and the pop single years

On 28 September 1955, between the Waller sessions and the Fitzgerald albums, Armstrong recorded Mack the Knife for Columbia, the Kurt Weill and Bertolt Brecht song from The Threepenny Opera that Lotte Lenya, Weill’s widow and the song’s original Jenny, had been promoting in America. Lenya attended the session, and Armstrong, characteristically, folded her presence into the record itself, calling out to her by name in the course of the performance as the song’s roll of characters unrolls, a piece of theatrical presence of mind that no arranger wrote for him. The record reached the American top twenty and became a standard of his stage act; Bobby Darin’s later version, in 1959, was the bigger hit, and Darin was frank about the source of his reading. The episode belongs in the catalogue for what it shows about the pop single years that followed. Armstrong in his fifties was a contender in the singles market on craft, taking a Weimar theatre song, a form as far from the race catalogue as the catalogue could hold, and making it swing without flattening its menace.

The later 1950s singles chart the same strategy across the market’s width: remakes of his own standards for new labels, film songs, the title performance for the film High Society in 1956, where his duet with Bing Crosby on Now You Has Jazz staged the meeting of the two dominant popular singers of the century as a friendly seminar on what jazz is. The singles were aimed at jukeboxes and radio, and some were product in the plain sense. But the catalogue rule this article has applied throughout holds here too: judge the playing. The trumpet on the Columbia and Decca singles of the period is rationed, as it had been since 1929, and what is there is placed with the old economy, and the voice, now unmistakably an instrument with a public personality attached, does the work the trumpet once did in spaces the trumpet no longer gets. The pop years also maintained the distribution machine, radio play, television appearances, film roles, that made the events of 1964 and 1967 possible. A number one single at sixty-two is not an accident that befalls a forgotten man; it is the yield of a decade of remaining unavoidable, and the remaining was work. The session itself shows the late method at work on unlikely material. The band was the working All Stars, Trummy Young, Barney Bigard, Billy Kyle, Arvell Shaw and Barrett Deems, and the arrangement leaves the verses to Armstrong’s voice and the choruses to his trumpet in the old rationing, except that the rationing is now the design rather than the constraint: the record is a vocal performance with instrumental commentary, and the commentary is unmistakably his. The song had a stage history that made his reading a small act of repossession. Written for Brecht and Weill’s Berlin premiere of 1928, carried to America in Marc Blitzstein’s English version, which had run off-Broadway with Lenya in the cast, Mack the Knife was European cabaret material, and Armstrong, the race catalogue’s graduate, took it into the American top twenty by treating it exactly as he had treated the Handy blues: as a song to be paraphrased, timed and sold by the force of the phrasing. Lenya’s presence at the session, and her reported delight at what he did with her husband’s music, closed a circle between two traditions of popular theatre that had never expected to meet.

The singles years also produced studio albums that the period division cannot file comfortably. Louis Armstrong Meets Oscar Peterson, recorded for Verve in 1957, is a quartet record of standards on which the trumpet and voice divide the programme between them with a Peterson trio in support, and it stands with the Fitzgerald albums as evidence of the decade’s studio seriousness. Louis and the Angels, for Decca in the same period, set the voice against strings in the manner of the era’s prestige pop albums, and Louis and the Good Book, in 1958, took the All Stars through a programme of gospel songs, the sacred repertoire of his New Orleans childhood recorded with the same conviction as the Saints record twenty years before. A catalogue in decline does not range like this. The pop single years were a portfolio, managed by Glaser’s office and executed by a musician who treated each format as a room to be played, and the rooms kept filling.

Hello Dolly and what it displaced

Did Armstrong really knock the Beatles off number one?

Yes. Armstrong recorded Hello, Dolly! in New York on 3 December 1963, and on the Billboard chart dated 9 May 1964 the record reached number one, ending a fourteen-week run in which the Beatles had held the top position with three successive songs. At sixty-two he became, at that date, the oldest performer to top the American singles chart.

The circumstances repay a catalogue’s attention. Hello, Dolly! was the title song of a Jerry Herman musical then approaching its Broadway opening, and the recording began as a demonstration record, cut for Kapp Records in December 1963 to promote the show to radio stations before the production opened. Armstrong’s All Stars, with Trummy Young’s successor Joe Darensbourg on clarinet in the group’s late configuration and Billy Kyle at the piano, played the tune as a piece of Dixieland showmanship, banjoist Danny Barcelona’s place in the rhythm section marking the group’s return to a New Orleans texture, and the record escaped its promotional purpose within weeks of the show’s January opening. Radio play built through the spring. The Beatles, who had occupied the American number one position without interruption since the first week of February, through I Want to Hold Your Hand, She Loves You and Can’t Buy Me Love, yielded the chart dated 9 May 1964 to a sixty-two-year-old singing a Broadway score. The album of the same name followed the single to the top of the album chart, the stage production’s fortunes rose with the record’s, and Armstrong collected the Grammy awards for the performance in the following year, including the award for the song itself as the year’s outstanding composition.

Here the ladder reaches its sixth and final rung, and the loop opened in section seventeen closes. Hello, Dolly! removed the Beatles from the top of the American chart in 1964, which makes the commercial peak of Armstrong’s recording career fall almost forty years after the records the canon is built on. Sit with the implication, because it is the sharpest single fact in the catalogue. The canonical story requires the peak of the man to sit in 1928 and everything after to be descent. The chart record, which is not a matter of taste, puts a second peak in 1964, higher in commercial terms than anything the young Armstrong achieved, in a market immeasurably larger than the race catalogue’s. How did a man in his sixties outsell the biggest band in the world? The answer assembled across this article is that there was no interruption to explain. The turn to popular song in 1929 had built the audience; the big band years had built the fame; the All Stars and the songbook albums had maintained the craft at full stretch; and when a Broadway tune met that apparatus, the result was a number one record that sounds, in its phrasing, its rhythmic poise and its trumpet choruses, like the work of the man who made West End Blues, because it was. The two peaks are one career, and the valley between them exists chiefly in the criticism. The Grammy honours of the following spring give the episode its official frame. Armstrong received the award for the year’s best male vocal performance for Hello, Dolly!, and Herman’s song took the song of the year prize, the industry’s establishment certifying what the chart had already said. The Kapp album built around the single extended the demonstration for home listeners: alongside the title song it carried his readings of A Kiss to Build a Dream On, the ballad he had first recorded in 1951 and now remade for the audience the single had won, Moon River, and a set of show tunes and standards sung and played with the All Stars’ late personnel. An album of new material by a man of sixty-two topping the American album chart was without precedent, and the precedent it set has not been repeated by any artist of his generation: the comeback narrative requires an absence, and Armstrong had never been absent, only misfiled.

The testimonials of the younger generation are part of the record’s documentation and part of this article’s evidence about reception. The Beatles themselves, asked about the displacement in the press conferences of that season, answered as fans rather than as victims, which they could afford to do and chose to do in the vocabulary of respect: the man who had taken their chart position was a musician their own musicianship had been schooled by, at one remove or another, and the trade papers printed the exchange as a coronation rather than an upset. A number one record is a commercial fact; a coronation by the displaced is a historical one, and the catalogue holds both.

What a Wonderful World and its afterlife

In August 1967 Armstrong recorded What a Wonderful World for ABC Records, a new song by Bob Thiele, the label’s producer, writing under the name George Douglas, and George David Weiss. Thiele had produced the session with a specific wager in mind: that the ballad method Armstrong had practised since Memories of You, applied to a song of unembarrassed sentiment over a string orchestra, could carry a single in a market then dominated by rock. The performance is the method in its purest late form. The trumpet is present chiefly as an introduction and a benediction; the record belongs to the voice, which delivers the song’s catalogue of ordinary wonders with a gravity that keeps the sentiment this side of syrup, the gravel and the timing doing the work that a younger singer would have asked of volume. In America the record failed on release, a failure the participants attributed to the label’s own ambivalence: Thiele recalled that ABC’s president, Larry Newton, disliked the record and declined to promote it, and the single stalled below the main chart. In Britain, issued in 1968, it climbed to number one and stayed there for a month, making Armstrong, at sixty-six, the oldest artist to top the British chart.

The afterlife of the record is the catalogue’s last lesson in how reception is manufactured. Two decades after its American failure, the song was placed in the film Good Morning, Vietnam, in a sequence that set its serenity against the war’s imagery, and the placement detonated a second career for the record: reissued in 1988, it reached the American top forty at last, and it passed into the small company of recordings that the general public, beyond any jazz audience, treats as permanent. The sequence of events, a flop in 1967, a British triumph in 1968, an American triumph in 1988, makes the record a controlled experiment in the argument of sections two and three. The performance did not change across those twenty years. The machinery around it, promotion, placement, fashion, did. A catalogue that can be misread that completely for that long, in both directions, by the market and by the critics, is a catalogue whose divisions deserve exactly the suspicion this article has applied to them. Armstrong lived to see the British success and to take the song into his stage act, where its sentiment, delivered by the century’s most unsentimental craftsman, reliably silenced rooms. He did not live to see the film that made it immortal in America, and the song’s posthumous rise belongs to the reception history of section thirty-two as much as to the record itself. The session’s construction is the ballad method’s purest surviving diagram. Thiele and Weiss wrote the song to Armstrong’s measure, its melody moving in the short, singable phrases his paraphrase technique was built on, and the arrangement, strings and choir in the era’s prestige manner, frames the voice the way the Russell band’s riffs once framed the trumpet: present, supportive, and never allowed to hurry him. His vocal takes the lyric at walking pace, weighting the ordinary nouns, trees, skies, babies, friends shaking hands, until the catalogue of commonplaces accumulates into something like a creed, and the trumpet’s brief appearance functions as an amen. Thiele’s account of the song’s origin, given in interviews across the years that followed, held that he wrote it for Armstrong specifically, having watched the man’s effect on audiences and on children in particular, and that the performance in the studio settled the room the way the record still settles rooms. Anecdotes from interested parties require care, and this article attributes the story as recollection rather than fact; but the record itself corroborates the tailoring. The song fits the method the way the Handy blues fit it, as if composed against a template the singer had spent forty years building.

The British success of 1968 deserves its own weighing, because it complicates any story of an American audience uniquely blind. Britain’s chart placed the record first for a month and kept Armstrong, who toured the country that year to enormous receptions, in the position of a reigning popular artist at sixty-six, and the British honours were followed by his celebrated British television appearances, in which the balladeer and the trumpeter shared the bill in his own person. The record’s Grammy Hall of Fame induction, in 1999, closed its institutional journey from flop to standard. Between the failure of 1967 and the induction, nothing in the grooves had changed; everything in the hearing had, and a catalogue article can offer the episode to its readers as the cleanest available parable of the difference between a record and its reception. The decline narrative, which this article has tested for thirty sections, is a reception artifact of exactly this kind, built from promotion, programming and period fashion, and the parable of What a Wonderful World is the form in which the whole argument can be carried away in a single song.

The late recordings and the physical decline

The final years of the catalogue were made against a failing body, and a works article should record the facts plainly, since the playing cannot be judged without them. Armstrong’s embouchure, the lip formation on which a trumpet career rests, had been damaged and repaired across decades of punishing schedules; he managed it with a discipline his sidemen described with awe, and the late records show the management as a changed distribution of effort, more voice, fewer stratospheric choruses, the climaxes built by placement rather than altitude. He had suffered a heart attack in Italy in 1959 and recovered to a full touring schedule. The recordings of the 1960s after Hello, Dolly! include a country music album, a set of Disney songs made in 1968, stage and film work, and two summit meetings with Duke Ellington in 1961, issued as Together for the First Time and The Great Summit, on which Ellington’s piano accompanies Armstrong’s readings of the Ellington songbook with a mutual deference that makes the albums a small genre of their own. The voice on the Ellington records, and on the Disney set, is the balladeer of 1930 perfected: nothing left but the phrasing and the warmth.

The catalogue closes with Louis Armstrong and His Friends, recorded in 1970 for the Flying Dutchman label, an album of gatherings and statements, among them a We Shall Overcome sung as a gathering hymn, and with a home recording of a spoken and sung seasonal piece he made for friends at the end of 1970. His last engagement, a two-week stand at the Waldorf Astoria’s Empire Room in February and March 1971, was completed in visible frailty and is documented on private tapes rather than commercial records. He died at his home in Corona, Queens, on 6 July 1971. The chronology of the life, the illness and the working years that produced these records is told in full in the companion account of Louis Armstrong’s life and career, and this article ends its own account where the catalogue ends: with a body of records that stops because the man stopped, not because the method failed. The last commercial sessions show an artist rationing a damaged instrument with complete intelligence, which is a different thing from decline, and the distinction is the one this article has been defending since its first section. The Ellington meetings of April 1961 deserve their paragraph in any account of the late work, because they test the method against the most sophisticated accompanist alive. Ellington, who had shared bills and mutual admiration with Armstrong since the 1920s, sat at the piano in a small band drawn from the All Stars, Trummy Young and Barney Bigard among them, Mort Herbert on bass and Danny Barcelona on drums, and played his own songbook behind Armstrong’s voice and trumpet for the Roulette label. The programme includes Duke’s Place, I’m Beginning to See the Light, and Azalea, a song Ellington had written for Armstrong years earlier and held back for the occasion. What the records document is deference in both directions: Ellington comps with a spareness his own band never heard from him, and Armstrong sings the Ellington melodies with a fidelity to the written line that his paraphrase method usually reserves for climaxes. Two catalogues, two methods, one afternoon’s understanding repeated across two albums, and the result belongs on any short list of his late masterpieces.

The management of the failing instrument is itself part of the late catalogue’s evidence. Armstrong’s lip had been a medical problem since the 1930s, when overwork in Europe left damage he managed for the rest of his life with salves, rest cures he rarely took, and a redistribution of his playing toward the middle register and the perfectly placed climax. A heart attack in Spoleto in 1959, during an Italian tour, hospitalised him and was followed, within weeks, by a return to the bandstand that his doctors had forbidden and his audiences demanded. Illness is not artistry, and this article does not make it so. But the late records’ economy, the sense that every note is spent where it counts, is partly the sound of a craftsman allocating a scarce resource, and listeners who call the late playing wise are responding to an allocation they can hear.

Reception then: what sold and what was reviewed

The contemporary reception of the records divides along the same line as their distribution. Inside the race catalogue, Armstrong’s records were commercial stock with a rising premium: the Hot Five sides sold steadily to the audience the series served, Heebie Jeebies was reported in the trade press as an exceptional seller for the catalogue, and the blues accompaniments moved with the fortunes of the singers they supported. Review coverage in the Black press treated him as a leading professional whose engagements and records were news. The white trade press discovered him in stages: the theatre work of 1929, the Cotton Club broadcasts from Los Angeles, and the films of the 1930s made him a personality before the jazz press, which barely existed at the start of the decade, had decided what he was. Down Beat, founded in 1934, and Metronome covered the big band as a going concern, reviewing the Decca records as popular product, sometimes warmly, sometimes with the qualification that would harden into the period division: admiration for the trumpet, condescension toward the programme.

The numbers tell the less mediated story. When the Saints Go Marching In in 1938 turned a sacred standard into a swing hit, the Decca record sold across the colour line the catalogues had enforced. The 1947 concerts played to the new concert audience that wartime and the LP would soon enlarge. The Columbia albums of the 1950s were prestige pressings reviewed in the general press as events. Hello, Dolly! was a gold record and a Grammy winner. What a Wonderful World sold in Britain while America ignored it, and the disparity is the period’s cleanest exhibit: the same performance, in the same season, a number one and a flop, the difference lying wholly in promotion and placement. Reception then, in short, was never a simple verdict on the playing. It was a set of markets with different gates, and the playing passed through every gate that was actually opened to it. The failures of reception in the catalogue are gate failures: records unpromoted, decades unreissued, a middle period unreviewed by critics who had decided its character in advance. A catalogue history that keeps that straight will not mistake the gates for the music, which is the mistake the two-career story institutionalised. The distribution of the early records deserves a word of its own, because it shaped what reception could even mean in the 1920s. Race records reached their audience through furniture shops, music dealers in Black neighbourhoods, mail order from newspaper advertisements, and, in the practice documented by the labels’ own salesmen, through Pullman porters who carried stock along the rail lines and sold it town to town. A Hot Five side could therefore be famous in places no reviewer ever visited, and the fame is real even though it left few paper traces. The Black press, the Chicago Defender and the Pittsburgh Courier chief among them, covered Armstrong’s engagements, his theatre appearances and his marital and managerial news as the doings of a star of their own world, years before the white dailies gave him comparable space. When the general audience arrived, by radio and film in the 1930s, it arrived to a reputation already a decade old, and the general press’s tendency to present him as a discovery says more about the press’s map than about his career.

Film compounded the effect. A musician whose records played under the dialogue of Pennies from Heaven in 1936, who appeared as a bandleader in Going Places in 1938 and in a procession of features thereafter, was being received by millions who never bought a jazz record and never read a jazz review. This matters to a catalogue article because it explains the otherwise puzzling shape of his commercial curve: the audience that bought Hello, Dolly! in 1964 had been accumulating since the talkies, through channels the jazz criticism of the day did not monitor and could not weigh. The critics were counting one audience, the connoisseurs of the 78, while the records were quietly building another, vastly larger one, whose loyalty would outlast every critical fashion. Reception since his death has been the story of those two audiences slowly becoming one, as the connoisseurs’ canon expanded to admit what the larger audience had valued all along.

Reception since: the canon and the reissues

Since his death the catalogue has been managed by the reissue industry, and the reissues have written the modern shape of his reputation more decisively than any single critic. The pattern began in his lifetime’s later decades, when French and American LP programmes of the 1950s and 1960s gathered the Hot Five and Hot Seven material into annotated sets, establishing the 1920s as the catalogue’s documented core while the Decca big band years circulated in unannotated budget pressings. The CD era corrected the fidelity and widened the net: the Mosaic label’s boxed sets, the Columbia and Legacy complete editions of the Hot Five and Hot Seven recordings, the JSP label’s chronological surveys that put the sideman work and the big band years back into print in sequence, and Decca’s own All Stars boxes. Each reissue programme is an argument about which Armstrong counts, and a reader comparing the programmes across forty years can watch the argument move: the early canon of the Okeh sides, the slow rehabilitation of the 1930s, the elevation of the 1950s Columbia albums to classic status, and the final absorption of the pop singles into the story as evidence rather than embarrassment. The 2000 television series Jazz, for which Giddins served as a senior adviser, carried the continuity reading to the largest audience the subject has ever had, and Riccardi’s archival work at the Armstrong House, opened as a museum in the house in Corona where the trumpeter lived, put the private tapes and papers behind the public records for scholars to test.

The lesson the reissue history teaches is the one this article has drawn throughout: the catalogue’s divisions track the availability and presentation of the records with embarrassing fidelity. Periods in print with annotation are periods with reputations. A listener assembling the evidence for a lecture, a broadcast or a classroom can lay those programmes side by side in a browser-based presentation viewer and watch the canonical Armstrong change shape decade by decade, which is a more instructive exercise than any single verdict on his periods. The records themselves, all of them, wait underneath the programmes, indifferent to the order in which they are issued, and the trend of fifty years of reissue work runs in one direction only: toward the whole catalogue, heard whole. Two honours mark the catalogue’s institutional standing and both are instructive. In 1977 the Voyager Golden Record, the phonograph carried out of the solar system aboard the Voyager probes, included Armstrong’s Melancholy Blues, the 1927 Hot Seven side on which he accompanies himself as much as he solos, chosen to represent the music of Earth to whatever might find it: the canon, at its most literal, selecting a race-catalogue performance as the human exemplar. The Grammy Hall of Fame, which inducts recordings of lasting qualitative or historical significance, has admitted Armstrong performances from across the catalogue, the early sides and the late ballads alike, and the Library of Congress’s National Recording Registry has likewise reached beyond the 1920s in its Armstrong selections. The honours matter less as laurels than as measurements: institutions charged with choosing a permanent record of American sound have consistently chosen more of Armstrong, and from more of his decades, than the two-career story would predict.

His own archive completed the correction from inside. Armstrong was a compulsive self-documenter who recorded his private hours on reel-to-reel tape, annotated his record collection, and kept the papers of a fifty-year career in the house in Corona that became the Louis Armstrong House Museum. Riccardi’s cataloguing and writings on that archive, and the museum’s programme of making the tapes and discographical materials available, gave scholars something no previous jazz generation had possessed: the artist’s own record of the records, against which the critical constructions can be checked. The private tapes show the connoisseurship behind the public catalogue, the musician at home studying his own past performances and the performances of others with the same analytical appetite the discographers brought to him. The archive does not settle arguments, but it changes their evidence, and the scholarship of Brothers and Teachout, working with the full documentary record, reads Armstrong’s career as a unity because the documents, unlike the reissue programmes, were never divided.

A listening order for a newcomer

Which Armstrong album should I start with?

Start with Louis Armstrong Plays W. C. Handy, the Columbia album of 1954. Its sound is full and modern, its programme is the blues songbook he grew up in, and its long performances show the mature method entire, the dialogue with the band, the building of a solo across choruses, the voice and the trumpet sharing one argument.

The order that follows is designed for a listener meeting the catalogue whole, and it proceeds by widening circles rather than by chronology, because chronology is the canonical division’s home ground. From the Handy album, move to the 1928 sessions: West End Blues, Weather Bird and the Hines sides, heard now with ears that already know where the method leads, so that the early records register as beginnings of a story rather than as a lost golden age. Then take the Hot Five and Hot Seven in sequence, the duet records and the Bessie Smith accompaniments, and only then the 1930s: Ain’t Misbehavin’ and Black and Blue first, the Decca ballads and the Saints after, hearing the big band as the popular theatre it was rather than as the decline it was advertised to be. Ella and Louis belongs in the middle of the journey, and the live albums, Symphony Hall above all, after it, since the concerts presume a listener who knows the repertoire they celebrate. Hello, Dolly! and What a Wonderful World come last, where their power to surprise is greatest: the show tune and the ballad heard as the final statements of a method built in 1925, which by then the listener can verify rather than merely believe.

Practical guidance belongs in a catalogue article. The compilations are legion and uneven, and section thirty-three’s companion answer in the questions below gives the navigation rule in brief: prefer the complete editions and the label’s own reissues over anonymous budget sets, and use the discographies, Morgenstern’s and Riccardi’s work chief among them, to place any track in its session. A listener keeping a dated listening log alongside the records, session by session, with the personnel and the label details noted as they go, will find the catalogue organising itself; such a log is well kept in the offline encrypted notes app, obtained from the VaultBook download page, where the notes stay private and searchable across the whole journey. What no order can supply is the one thing the records keep offering on their own: the experience of hearing the supposed second career outplay its legend, which is the experience this article has been documenting from its first page. A second pass, for the listener the first journey has converted, can afford the chronological route the first pass avoided, and it will sound different from the textbooks’ version of chronology. Heard in sequence after the whole catalogue is known, the Oliver breaks are no longer a prologue but a first chapter of a continuous narrative; the big band records, encountered in their session order with the discography open, reveal their own internal development, the band tightening through the later 1930s, the ballad method deepening year by year; and the 1950s albums arrive not as a comeback but as a consolidation. The newcomer should also be warned about the catalogue’s traps. European public-domain reissues of variable transfer quality can make the early records sound worse than they are; multi-artist compilations pad Armstrong’s name across sides on which he plays two breaks; and session credits on budget sets repeat errors the scholarly discographies corrected decades ago. The antidote is the session habit: any Armstrong record worth hearing is worth hearing with its date, its label, its personnel and its place in the sequence known, which is how the records were made and how they make sense.

One more category belongs in the order, and it is the live one. After the studio journey, hear the 1947 Symphony Hall concert and the 1956 stadium performance preserved in Satchmo the Great in full, because the concert Armstrong is the control against which the studio rationing becomes audible: given a hall, an occasion and no three-minute wall, the method expands to fill it, and the listener who has worked through the catalogue will hear forty years of records being summarised by their maker in real time. It is the nearest thing the catalogue offers to the artist’s own commentary on itself.

Louis Armstrong Greatest Recordings: One technique, two markets

The catalogue, followed in order from the Gennett studio in 1923 to the last sessions of 1970, yields a single conclusion with two halves. The first half is about markets. Armstrong recorded for a segregated race catalogue, for the theatre, for the dance band circuit, for the European concert stage, for the LP songbook, for the duet album, for the Broadway tie-in and for the pop single, and every one of those markets asked him for a different product and rationed his gifts differently. The second half is about technique, and it is the half the canonical division gets wrong. The technique did not change hands, retire or decay. A lead part conceived as a paraphrase that intensifies; a sense of time that organises every band it enters; a dialogue habit that turns accompaniment, duet and self-duet into one continuous practice; a solo architecture of statement, development and climax fitted to whatever space the format grants: these are present in the Oliver breaks, in the Smith accompaniments, in the Hot Five, in the Hines dialogues, in the big band choruses, in the Handy choruses, beside Ella Fitzgerald, and inside a two-minute Broadway single cut by a man of sixty-two. The evidence for continuity is not a mood or a loyalty. It is the records, in sequence, and the sequence does not break.

The structural close this article owes its reader is the test that would settle the remaining dispute. For the decline narrative to be right, there would have to be a measurable failure in the playing itself, a date after which the time, the tone, the invention or the architecture gave way, visible in the records the way the 1928 transformation is visible. Fifty years of reissues have made the whole catalogue available for that inspection, and the failure does not appear: what appears instead is rationing, the formats granting less space to the horn as the voice’s market value rose, and an artist managing a damaged lip and an ageing body with undiminished intelligence about where a climax belongs. For the continuity case to be wrong, the late records that this article has rated at the first rank, the Handy and Waller albums, the Fitzgerald duets, the ballad singles, would have to be reclassified as nostalgia, and they resist the classification on every technical ground the canon itself uses. The title of this closing section names the finding in its plainest form. Two markets is not a metaphor. For the first half of the catalogue the markets were literal and enforced: a race catalogue that sold the Hot Five to one America and a general market that could not be entered until the theatre, the broadcasts and the films pried it open. For the second half the division ran between the connoisseur market, which bought the catalogue as art and policed its periods, and the popular market, which bought the singer, the personality and the singles. Armstrong served all of these markets without changing the goods. The paraphrase, the time, the dialogue instinct and the architecture are the goods, and a shopkeeper who keeps the same stock for forty-eight years while the street outside changes beyond recognition is not two shopkeepers. The two-career story persists in reference works partly because it is tidy and partly because period divisions organise reissue programmes and syllabus weeks, but tidiness is not evidence, and the syllabus that splits him at 1929 teaches its students to hear a break that the records themselves decline to perform.

What remains, when the division is set aside, is the largest single body of foundational recordings in American music, made by one musician whose method was complete by the time he was twenty-seven and capacious enough to hold everything the century’s markets asked of it. The catalogue asks to be heard whole, and heard whole it coheres.

Frequently Asked Questions

Q: What are Armstrong’s most important recordings?

The essential shelf begins with the records of 1928: West End Blues, with its opening cadenza and its blues choruses, and Weather Bird, the duet with Earl Hines that reduces his art to two musicians in conversation. Around them stand the Hot Five and Hot Seven sessions of 1925 to 1927, Cornet Chop Suey, Potato Head Blues and the blues accompaniments, above all St. Louis Blues with Bessie Smith. The big band years contribute Ain’t Misbehavin’, Black and Blue and When the Saints Go Marching In. The LP era adds Louis Armstrong Plays W. C. Handy, Satch Plays Fats, the Ella Fitzgerald duet albums and the Symphony Hall concert of 1947. The popular peaks, Mack the Knife, Hello, Dolly! and What a Wonderful World, complete the shelf, because the catalogue’s argument is that all of these are one body of work.

Q: What is the difference between the Hot Five and the Hot Seven?

The Hot Five was the Chicago studio group of 1925 to 1927, Armstrong with Kid Ory on trombone, Johnny Dodds on clarinet, Lil Hardin Armstrong on piano and Johnny St. Cyr on banjo, a New Orleans front line without drums. The Hot Seven was the same group expanded for one week in May 1927, with Pete Briggs on tuba and Baby Dodds on drums added, and John Thomas deputising on trombone. The expansion changed the music’s weight and freedom: the rhythm section let the tempo breathe, the solos lengthen, and the slow records, Willie the Weeper and the Wild Man Blues among them, gain their spaciousness from the added chairs. The personnel and the studio circumstances were otherwise continuous, the same label, the same city and the same leader, and the two names describe one band at two sizes rather than two bands.

Q: How were the Hot Five sessions recorded?

The sessions were made for OKeh’s race catalogue at the label’s Chicago studio, beginning with electrical recording after a single acoustic date in November 1925. The band was not a working group: its members held separate nightly jobs, Lil Hardin Armstrong leading her own band across town, and it assembled only in the studio, reading Lil’s arrangements and the standard New Orleans book by ear and memory. Sessions produced two to four titles, generally in a pair of takes each, with the issued master chosen by the label’s supervisor. The records were marketed in the race series, sold through the segregated distribution of the period, and credited to a name, Louis Armstrong and His Hot Five, that existed nowhere outside the studio and the catalogue. Surviving alternate takes show the solos rebuilt between takes, evidence that the performances were improvised within prepared designs.

Q: Who played on the 1928 Earl Hines sessions?

The band was Carroll Dickerson’s Sunset Cafe orchestra contracted to a studio group, billed on the labels as the Savoy Ballroom Five. The core personnel across the June and July sessions were Armstrong on trumpet, Fred Robinson on trombone, Jimmy Strong on clarinet and tenor saxophone, Earl Hines on piano, Mancy Carr on banjo and Zutty Singleton on drums. Hines’s presence is the sessions’ glory: his piano, played in a horn-derived style, made the records dialogues between equals, in West End Blues, Fireworks and Muggles. Discographies have debated details of the December sessions’ personnel, particularly the reed and drum chairs on individual titles, and the responsible course in any sessionography is to give the standard listing and flag the debated titles rather than present either version as certain. The sound, whoever precisely sat in each December chair, is unmistakably the same musical organisation.

Q: What makes the opening of West End Blues famous?

The record opens with an unaccompanied trumpet cadenza of some fifteen seconds, a cascade of perfectly weighted runs that Armstrong plays alone before the band enters. Its fame rests on several grounds at once. As pure playing, its control of time and tone has never been surpassed on record, and trumpet students learn it by ear as a foundation text. As composition, it states the blues’ whole expressive range in miniature, the flourish, the cry and the poise held in one gesture. As history, it announced the soloist’s era: after 1928 the improvised solo, not the ensemble, was jazz’s centre, and this cadenza is the announcement preserved in sound. It is also perfectly placed, a curtain-raiser that the following choruses justify, so that the record plays as a single rising argument from its first note to its vocal close.

Q: Why did Armstrong accompany blues singers early on?

Accompanying singers was the steadiest studio work available to a young cornetist in the race catalogue economy of 1923 to 1925, and Armstrong was in demand for it because the work suited gifts he had in unusual measure. A singer’s accompanist must support without competing, answer phrases in the gaps, and build an ending, and those obligations trained the dialogue method that governs his whole catalogue. The singers were also the market’s stars, Ma Rainey, Bessie Smith and the rest, and their sessions were prestige dates on which a sideman’s sound reached the widest audience the catalogue had. The artistic reason completes the economic one: the blues singers were the music’s deepest school of phrasing and time, and the young Armstrong, decorating their choruses, was being educated by the best teachers in American music. His later singing applies their lessons in his own voice.

Q: Why did Armstrong move to popular songs?

The move was the market’s before it was the musician’s. By 1929 the race catalogue’s centre was shifting from instrumental jazz to songs, the theatre offered the general audience, and a singing bandleader could reach buyers that a trumpet feature could not. Armstrong’s New York theatre work, in Hot Chocolates and Connie’s Immortals, put Ain’t Misbehavin’ and Black and Blue into his book, and the records followed the stage. He also had the instrument for the turn: a voice whose phrasing applied his trumpet method to a lyric, bending syllables to the beat and paraphrasing melodies exactly as the horn paraphrased them. The songs widened his fame without retiring the horn, which the same records show rationed rather than diminished, its choruses placed for effect. The turn looks like a fall only under the assumption that popular song is a lesser material, an assumption the performances themselves refute.

Q: Did critics reject Armstrong’s big band recordings?

The rejection was real but narrower than its legend. When a specialist jazz press formed in the 1930s, its first critics consecrated the 1920s records and treated the big band Decca sides as commercial product, sometimes warmly reviewed as entertainment, rarely weighed as art. The general press and the Black press judged differently, covering the band as a leading attraction, and the records sold across the decade to an audience that had not consulted the specialists. Later critics, Schuller in 1968 and Giddins across his career, reassessed the decade from the records and found the judgement wanting: the ballad performances, the remakes of his own classics and the best Decca sides show the method intact inside a rationing format. The honest summary is that a generation of tastemakers rejected the records on principle, the public did not, and the records have outlasted the principle.

Q: Are the All Stars records considered good?

Yes, and the judgement has strengthened with time. The All Stars, formed in 1947 after the big band’s collapse, returned Armstrong to a front-line format with masters beside him, Jack Teagarden, then Trummy Young, Barney Bigard, then Edmond Hall, and the group’s records, the Symphony Hall and Pasadena concerts, the Decca and Columbia studio work, are now rated among his essential documents. Contemporary critics of the traditionalist persuasion praised the band while some modernists dismissed it as a museum act, but the recorded evidence, the Handy and Waller albums above all, settled the question in the band’s favour. The records considered weakest are the early 1950s jukebox singles, where the format rationed him hardest. Judged by the concerts and the songbook albums, the All Stars stand as the equal of any group he led, the 1928 settlement restored at chamber scale.

Q: What is the Plays W C Handy album?

Louis Armstrong Plays W. C. Handy is the Columbia album recorded in July 1954 with the All Stars, produced by George Avakian, devoted entire to the published blues of W. C. Handy. Its programme, St. Louis Blues in a version past nine minutes, Beale Street Blues, Memphis Blues, Atlanta Blues, Yellow Dog Blues and others, treats Handy’s book as a songbook, the way a classical singer treats a composer, and the performances give the mature Armstrong room the three-minute era never allowed: long trumpet statements, dialogues with Trummy Young’s trombone and Barney Bigard’s clarinet, and vocals shared with Velma Middleton. Handy, elderly and blind, heard the record and approved it, and the album has been rated since its release as the summit of the late catalogue, the strongest single evidence that the technique of 1928 survived, deepened, into the LP age.

Q: How good are the Ella Fitzgerald collaborations?

They are among the finest vocal records either artist made. Ella and Louis, recorded for Verve in 1956 with the Oscar Peterson quartet, sets the two voices in duet across the standard songbook, her purity against his grain, his trumpet answering both, and the album’s ease conceals a perfect technical accommodation between opposite methods. Ella and Louis Again, a double album from 1957, extends the programme, and Porgy and Bess, from the same year with Russell Garcia’s orchestra, carries the partnership into Gershwin’s opera score. Contemporary reviewers treated the sessions as a summit, and posterity has agreed: the records appear on every serious list of the era’s vocal albums. For Armstrong’s catalogue they prove a specific point, that beside the most accomplished singer of the age, in the most exposed format in jazz, his singing and playing lose nothing by the comparison.

Q: Did Hello Dolly reach number one?

Yes. Armstrong recorded the song in New York on 3 December 1963 for Kapp Records, originally as a demonstration record to promote Jerry Herman’s approaching Broadway musical. Radio play built through the spring of 1964 as the show opened, and on the Billboard chart dated 9 May 1964 the record reached number one, displacing the Beatles, who had held the position for fourteen consecutive weeks with three successive records. At sixty-two Armstrong became, at that date, the oldest artist to top the American singles chart. The album of the same name also reached number one, and the performance won Grammy honours the following year. The episode is the catalogue’s sharpest commercial fact: the peak of his chart career arrived almost forty years after the Hot Five, achieved by the same phrasing and time that the canonical records display.

Q: Should I listen to the studio or the live records first?

Begin with the studio records for the map, then take the concerts for the territory. The studio catalogue, the Hot Five and Hot Seven sessions, the Handy album, the Fitzgerald duets, gives the method in its most concentrated and best-recorded forms, and a listener who knows those records can hear what the concerts are doing. The live documents, the 1947 Symphony Hall concert above all, then reveal the proportions of a working evening, the trumpet’s true share of the night, the pacing, the features for the sidemen, that the three-minute studio era conceals. Reversing the order has its own pleasure, the concerts’ generosity first and the studio’s intensity after, but the map-first route protects a newcomer from the catalogue’s one real difficulty, its sheer size. Ninety records into either route, the two bodies of work explain each other, and the distinction stops mattering.

Q: Is it worth hearing the late recordings?

Emphatically, because the late catalogue contains some of the strongest evidence in his whole body of work. The Handy and Waller songbook albums of 1954 and 1955, the Fitzgerald duets, the Oscar Peterson quartet album, the Ellington summit meetings of 1961 and the best All Stars concerts stand at the first rank by the same technical measures applied to 1928. Even the pop records reward the ear: Mack the Knife’s theatrical timing, Hello, Dolly!’s phrasing, and What a Wonderful World’s ballad gravity are performances by a master allocating a rationed instrument with complete intelligence. The very last sessions show frailty managed rather than hidden, and they close the catalogue with dignity. A listener who stops at 1930, obeying the old period division, keeps the foundation and throws away the house, including the records by which the foundation’s meaning is confirmed.

Q: How do I navigate the many Armstrong compilations?

Prefer completeness and provenance to price and bulk. The safest routes are the labels’ own complete editions, the Columbia and Legacy sets for the Hot Five and Hot Seven years, the Mosaic boxed sets for the studio sessions they cover, the JSP chronological surveys for the sideman and big band periods, and Decca’s own reissues for the All Stars. Treat anonymous budget compilations with suspicion: their transfers are often poor, their session credits repeat old errors, and multi-artist sets may pad his name over sides where he plays two breaks. Keep a session reference open as you listen, the printed discographies and the Armstrong House’s materials, and log each record with its date, label and personnel. The catalogue organises itself quickly once every track has a session attached to it, and the compilations become a library rather than a maze.

Q: How do I identify which session a track comes from?

Armstrong’s style changed audibly across the decades, and the ear learns the markers. The 1920s sides are small-group records, acoustic or early electrical, with the trumpet brilliant and high and the vibrato narrow. The 1930s records carry a big band frame and a mellowing trumpet featured in choruses, with the voice assuming half the programme. The All Stars records are unmistakable, the front line of trombone and clarinet around him, and the LP-era studio sound is warmer and longer. When the ear hesitates, the discographies decide: the standard sessionographies, Morgenstern’s and Riccardi’s work among them, list every session with date, label, personnel and matrix numbers, and a track’s matrix or catalogue number, printed on any responsible reissue, keys it to its session exactly. Uncredited appearances, the blues accompaniments above all, remain partly matters of expert attribution, and honest references flag those.

Q: Which five Armstrong recordings give the best overview?

First, West End Blues of 1928, for the cadenza and the complete solo architecture in a single side. Second, St. Louis Blues of 1925 with Bessie Smith, for the accompanist’s art that founded everything. Third, Ain’t Misbehavin’, in either the 1929 theatre version or the Waller album remake, for the popular-song method at its most complete. Fourth, St. Louis Blues again, this time the 1954 Handy album version, nine minutes of the mature method proving itself against the young man’s signature material. Fifth, What a Wonderful World, for the ballad gravity of the final manner and the catalogue’s last word on reception. The five span forty-two years, four formats and both markets, and they make the article’s argument in miniature: one technique, developing, across the whole of a working life.

Q: What should I listen for in a Hot Five record?

Listen first for the allocation of space. The leader rations himself, accompanying, answering and building toward the choruses that matter, so that the trumpet’s entrances feel like events. Listen for the dialogue, the clarinet and trombone commenting around the lead, and for the way the lead comments in return when others solo, the accompanist’s art practised inside his own band. Listen for the architecture of his solos, a statement near the melody, a development that paraphrases it, a climax placed where the side’s remaining time can hold it. Listen for the time, the beat’s weight shifting as the band moves from ensemble to feature. And listen for the humour as musicianship, the spoken asides and novelty effects timed like breaks, because the comedy and the virtuosity are one professionalism, the sound of a repertory company that knew exactly what it was doing.

Q: How did reception of Armstrong’s later records change?

The change ran from condescension through embarrassment to canonisation. In his lifetime the specialist press treated the post-1930 records as the entertainment of a great man in decline, a judgement the general audience ignored while buying the singles and the albums. The first scholarly reassessments, Schuller’s among them in 1968, began the correction, and the reissue era completed it: as the Handy album, the Fitzgerald duets and the big band boxes returned to print with annotation, critics working from the records rather than the legend re-rated the late catalogue upward decade by decade. The television series Jazz in 2000 and the museum scholarship of the Armstrong House carried the continuity reading to a general audience. The later records are now routinely described as masterpieces in the reference literature, a verdict the performances had earned in 1954 and waited forty years to collect.

Q: How did reissues shape which records are remembered?

Reissues decided the catalogue’s public shape, because a record out of print is a record out of the argument. The 1920s sides were reissued early, lovingly and with annotation, first by the French collectors’ labels and then in LP programmes, and they became the documented canon. The 1930s Decca records circulated in unannotated budget pressings and were judged, when judged at all, from a corrupted sample. The CD era corrected the balance: complete editions restored the big band years to sequence, the Mosaic and Legacy boxes gave the studio sessions scholarly editions, and the All Stars albums stayed in print until their standing was undeniable. Each generation’s Armstrong is therefore partly an artifact of its reissue programme, and the direction of fifty years is consistent, toward more of the catalogue, better documented, with the period divisions weakening as availability equalises. The records remembered best are the records kept available.