Singing Zarzuela, 1869–1958

Summary

In the last few years, digitizations and reissues of historical recordings of Spanish zarzuela - from wax cylinders in the 1890s to long-play records in the 1950s - have revealed a range of contrasting vocal performance styles. By focusing on portamento, this Element sets the foundations for a contextually sensitive history of vocal performance practices in zarzuela. It takes stock of technological changes and shifts in commercial strategies and listening habits to reveal what the recorded evidence tells us about the historical development of portamento practices and considers how these findings can allow us to reconstruct the expressive code of zarzuela as it was performed in the late nineteenth century and how it transformed itself throughout the next half century. These transformations are contextualized alongside other changes, including the make-up of audiences, the discourses about the genre's connection to national identity and the influence of other musical-theatrical genres and languages.

Introduction

When listening to recordings of vocal and instrumental works made around 1900, our attention is likely to be drawn, for better or worse, to portamento. Instances of portamento in these recordings are indeed considerably greater than we would hear in present-day performances of the same works. From our modern perspective, we may regard these earlier interpretations as being overly sentimental or kitsch. Zarzuela, the Spanish-language musical theatre genre that flourished between 1850 and 1950, was no exception.¹ This becomes obvious from even cursory listening to the increasing number of historical recordings that have become widely available in recent years, both through digitization projects led by research libraries² and through commercial CD releases.³

The appearance of this wealth of material has the potential to transform the ways in which we study and understand zarzuela, including how it was interwoven into its cultural and social contexts and assessing its place within broader styles of singing and their developments between the late nineteenth and mid twentieth centuries. Scholarship on zarzuela has grown significantly since the 1980s, after musicology established itself as a discipline in its own right within Spanish academia and zarzuela became one of its pre-eminent objects of study. Out of necessity, much of the work conducted in the initial decades was foundational and was led to a great extent by the Instituto Complutense de Ciencias Musicales (Complutense Institute of Musical Sciences) at the Universidad Complutense de Madrid. Staff at t he Instituto compiled an ambitious dictionary (Diccionario de la zarzuela en España e Hispanoamérica) that included entries on composers, works and singers,⁴ published biographies of the main composers of the genre⁵ and launched a series of editions that included some of the best-known repertoire works (Colección Música Hispánica – Música lírica). At a later stage – and more within the field of Spanish cultural studies rather than musicology – new research directions pushed the study of the genre beyond the ‘composers-and-work’ paradigm and focused on its cultural, social, political and ideological dimensions, such as the place of zarzuela within discourses of national identity and studies on venues and audiences.⁶

The realm of zarzuela performance, however, has remained surprisingly unexplored. Studies of the social and cultural history of performers⁷ and particularly of performance practice remain limited.⁸ This is somewhat paradoxical. Presumably, a main reason why audiences kept filling zarzuela theatres was because of the fascination that the music, as conveyed in the act of performance, exerted on them. Yet we know little about the sonic means through which audiences perceived zarzuelas and their associated discourses and connotations.

Against this backdrop, this Element intends to set the foundations for a contextually sensitive history of vocal performance practices in recorded zarzuela by tracing the story, from the late nineteenth to the mid-twentieth centuries, of one of the most conspicuous features heard in these early recordings: portamento. The start and finish dates frame a cohesive unit within the history of recorded music in Spain. In 1896, the first commercial recordings of zarzuela in Spain are likely to have been made and in 1958, Ataúlfo Argenta, a prolific conductor who pioneered the recording of complete zarzuelas on LPs (long-playing 33 rpm records), passed away. While zarzuelas certainly continued to be recorded after 1958, Argenta’s death marked, in a way, the end of an era. The 1950s also brought fundamental changes to the genre’s popularity on stage. Once a thriving art form filled with premieres of new works, by the end of the decade zarzuela had become increasingly fossilized into a set repertoire of canonic works, a process to which Argenta’s extensive catalogue contributed considerably.

Portamento is certainly not the only aspect of performance in earlier zarzuela that merits attention, but it is particularly important because it offers a compelling way to hear the interweaving of performance style with the genre’s social and cultural contexts. This can allow us to explore how and why styles can change according to social and cultural shifts without resorting to simplistic cause-effect explanations. One particular aspect of portamento that allows us to combine performance practice with contextual study in this way is that it was an aspect of performance that zarzuela shared with contemporaneous recorded genres, including opera, art song, various vernacular vocal styles and a wealth of instrumental music. This necessitates that we engage with questions such as how zarzuela interfaced and interacted with these other genres, both in Spain and internationally, how these interactions were regarded by composers, musicians, critics and audiences at different points in time and, ultimately, how the genre was perceived to represent national identity in a distinctive way. These are indeed key questions in zarzuela scholarship that go beyond performance. While the genre was universally recognized as a distinctively Spanish art form, a range of controversies and debates emerged throughout the decades on the extent to which foreign influences were to be either celebrated or decried.⁹

While some of these controversies and debates will be spelled out in more detail throughout this Element, there are two general themes in the reception of zarzuela in Spain that I wish to call attention to at this point, for they provide a foundation for understanding much of the history of the genre, particularly to newcomers. First, the discourses of Spanish identity upon which zarzuela was built and which it disseminated to its audiences were not monolithic. Different notions of Spanishness coexist in zarzuela and perceptions and debates concerning Spanish national identity changed throughout the period under study. Second, controversies abound surrounding the relationship between zarzuela and opera. While the two genres obviously share a great deal in terms of compositional techniques and singing styles, some in the Spanish music scene at different points in time have regarded zarzuela as inimical to opera, arguing that Spanish composers had been drawn to zarzuela because it was popular and profitable. By writing zarzuela, it was thought that they were neglecting to build a tradition of Spanish opera, one that would allegedly have given Spanish music a more universal appeal.¹⁰

Apart from disentangling the complex threads of Spanish national identity, zarzuela and opera, focusing on the study of recorded portamento also allows us to examine how performance practice was interwoven into the history of recording technologies and indeed into the recording industry itself. In order to explore the interface between individual performance practices and the relevant technological conditions, this Element, rather than focusing on a small number of well-known performers (as has often been the case with earlier studies of performance practice as reflected in early recordings), engages with an ample number of singers from various backgrounds. These recordings often reflect what a majority of contemporary zarzuela critics, audience members and singers may have considered less-than-optimal performance decisions. Such choices, as we will see, can often be chalked down to a lack of experience or expertise on the part of the recordist or of the singer, particularly in the latter’s dealings with technology.¹¹ Rather than simply dismissing these moments as unrepresentative, this Element regards them as opportunities to consider how Spanish recordists and singers interacted and experimented with the parade of successive new technologies arriving in the country. These interfaces and experiments sometimes influenced performance practices on stage in what Mark Katz has termed the ‘phonograph effect’. In such instances, recording technologies do not simply document stage performance practices but rather reflect changes in performing styles that happened in the recording studio and subsequently became commonplace in live performance.¹²

Beyond looking at the impact of recording technologies on performance practice, focusing on the early recording industry also allows the exploration of related topics that go beyond the merely technological, including, for example, the impact of recording technologies on the music profession as well as on listening and consumption practices.¹³ The formats and conventions we hear in early recordings – particularly those that might seem more alien to us – also allow us to consider broader questions pertaining to ontologies of recorded music vis-à-vis live performance, listening practices and the circulation of recorded music. In turn, we can analyse how these dimensions may have panned out in distinct ways within the context of Spain and zarzuela, thus contributing to the shaping of discourses surrounding national identity. However, while zarzuela enjoyed a significant degree of circulation in the Americas (in both live and recorded formats), recordings were primarily made in Spain. (Exceptions will be noted and discussed in due course.) Focusing on Spanish-made zarzuela recordings facilitates the comparison between recordings that essentially came from the same milieu and which were closely connected to the genre’s vibrant live culture, at least until its decline in popularity.

Interpreting Portamento through Zarzuela Recordings

This Element takes as its starting point the fact that portamento in zarzuela did not develop in isolation from other vocal genres. Of these, opera and art song are the only ones that have thus far received sustained scholarly attention and this scholarship informs the present enquiry. The connection between zarzuela and other sung genres is confirmed by historical evidence; indeed, the origins of modern zarzuela in the mid nineteenth century were decisively influenced by Italian bel canto and French opéra comique and opérette, not just in terms of compositional language but also performance practice. Zarzuela singers at that time trained either at the Real Conservatorio de Madrid or privately with opera-focused singing teachers, many of whom were Italian.¹⁴ Throughout the history of the genre, it was not unheard of for singers to move back and forth between zarzuela and opera, which suggests that singing practices were thought to be transferable, at least to an extent, between the genres.

Evidence suggests, however, that performance practices, though transferable, might not have been thought to be exactly equivalent, for zarzuela could also accommodate numbers with more traditional or popular streaks. These songs presumably called for performing styles more akin to those heard in music hall, cabaret and similar urban genres that were being recorded elsewhere at the turn of the century. Such vocal deliveries were characterized by a low soft palate that resulted in a nasal tone and intelligible diction, both of which were key to making humorous texts intelligible.

Similarly, when leading opera singers ventured into zarzuela, their approaches to singing could sometimes raise eyebrows. The Teatro Real in Madrid, although specializing in opera, occasionally hosted one-off performances of successful zarzuelas outside the main season. Some roles went to well-known zarzuela performers and others to operatic celebrities. In 1904, Ruperto Chapí’s La revoltosa was staged with opera stars Matilde Lerma, Andrés Perelló de Segurola, Luisa García Rubio and Emilio Cabello in the main roles. Only one zarzuela performer appeared with them, the baritone Emilio Mesejo, who at the time was immensely popular. Reviews of these performances at the Real document certain performance practices that we do not normally hear in recordings or find in other reviews. Lerma allegedly interpolated a high C in a duet that was said to ‘leave her partner, Mesejo, speechless and shocked’.¹⁵ Zarzuela was seen as being able to integrate such hybrid elements into itself, unlike opera.

The definition of portamento employed in this Element is deliberately a simple one: following Ellen T. Harris in The Grove Dictionary of Music and Musicians, it can be thought of as ‘the connection of two notes passing audibly through the intervening pitches’.¹⁶ This definition is well suited to the period under study and aptly describes what we hear in zarzuela recordings. While in the eighteenth century and earlier the notion of portamento di voce was broad and vague, by the early nineteenth century its meaning had been restricted to the expressive sliding which is familiar to us today.¹⁷ Within zarzuela recordings – as well as in contemporary examples from other genres – portamento reflects some variability, for example, in terms of speed or the size of the intervals to which it was applied. This Element will examine how this variability maps out over time and what factors might have shaped its different directions.

Several conclusions drawn from existing scholarship on portamento inform my approach to zarzuela recordings. First, a key preoccupation in many studies has been to deduce from existing recordings the rules that might have determined when and how portamento should have or could have been applied.¹⁸ Written sources, such as treatises and reviews, do not always help in this regard. Sometimes they may warn against performing portamento in ways that go against good taste, but rarely do they define what is meant by good taste. This is seen in Spanish sources from the period at hand¹⁹ as well as in others from throughout Europe.²⁰ Our modern ears, unaccustomed as they typically are to portamento, might be unreliable in determining what constituted good taste at the time and what did not. On the basis of recordings, a range of literature has convincingly mapped out portamento conventions, for example, within the boundaries of a composer’s output,²¹ a single song²² or a particular country.²³ My approach here follows a similar logic: given that portamento conventions were not spelled out – at least in writing – our best means to understand how they worked is to compare a manageable body of recordings that share a common notion of style. While outliers and unorthodox practices certainly made their way into recordings, such comparisons can help us identify norms and exceptions and detailed examinations of the historical contexts can assist in understanding how these practices came about and what they reveal about the relationships of singers, listeners and recordists to technology.

Another result of earlier studies relevant to the present one concerns the relationship between portamento and legato. Scholars have repeatedly pointed out that portamento is not only a deliberate expressive choice in musically or dramatically important moments but also can operate as a technical aid to facilitate legato (smooth) motion on difficult or large intervals.²⁴ Spanish primary sources explicitly recognize that portamento had this dual nature. José Aranguren’s Prontuario del cantante e instrumentista (1860) states that portamento could be used merely as an aide to achieve legato between syllables, though it could be exaggerated and thus become an intentional expressive device.²⁵ The legato function of portamento can be regarded as part of what Roberta Plack has labelled ‘vocal habits’, that is, the default way in which a singer approaches her instrument before adding any intentional expressive affects.²⁶ A singer’s vocal habits, or defaults, are determined by technical and physiological factors (e.g., a singer’s age and her ability to move from one pitch to another with or without having to resort to audible portamento) as well as conventions of technique and style in fashion at any given time.²⁷ Again, comparisons between recordings and examinations of portamento within the context of other aspects of vocal performance are useful in clarifying where the boundary might lie between vocal habit and deliberate expressiveness and how this boundary has shifted over time.

Explorations of the connection between portamento and legato has also led scholars to question the broader meaning and role of portamento. In particular, several researchers have suggested that there is a connection between portamento and speech; this is an avenue that I also intend to pursue throughout this Element. In situating portamento as an integral part of vocalism, Deborah Kauffman was the first to suggest that portamento might have had the general aim of making singing more reminiscent of speaking.²⁸ In the final section of John Potter’s article about portamento, which analyses recordings by Adelina Patti, Nellie Melba and other opera singers, the author speculates that ‘portamento helped give an illusion of language, re-creating the contour (as opposed to the sound) of speech in exaggerated form’.²⁹ Daniel Leech-Wilkinson, in an article that is more speculative and exploratory than purely empirical, suggests that portamento can be seen as an example of the more ‘emotionally straightforward’ nature of musical performance before the Second World War,³⁰ one that appealed to older audiences because it mimicked their earliest, most instinctive emotional responses to musical sounds.³¹

Primary sources suggest that this connection between portamento and language can be productive in the study of performance practice in zarzuela, while at the same time illuminating features unique to zarzuela that are not present in other genres. As discussed above, written primary sources are not always terribly helpful in clarifying the attributes of a good zarzuela performance (as opposed to opera) and likewise the overarching principles that governed the use of portamento. The singing treatises that were in use at the conservatoires and in private singing studios throughout the period 1850–1950 often point out that portamento should be used with discernment and good taste, but without explaining in detail what this actually meant. This aligns with singing treatises elsewhere in Europe, as discussed above.³² These treatises are also not specific to zarzuela and so they do not offer hints as to what makes a good performance in this particular genre.

Reviews published in newspapers also tend to provide little detail about specific aspects of performance. Singers’ performances were usually described with just a few words or, at best, a couple of sentences. General adjectives (e.g., ‘graciosa’ (graceful)) were used to tell about the performer, without any detailed discussions of specific stylistic elements. Specialized music periodicals often provided more detailed commentary, but writers here tended to focus on matters of composition and the troubled relationship between zarzuela and Spanish opera or on general news and gossip about musical life and zarzuela companies.

There is, however, a constant that runs throughout the zarzuela-focused musical press and music criticism that can provide a guiding principle to evaluate portamento (and other performance elements) in historical recordings: namely, the idea that good singing in zarzuela equated good, expressive and communicative singing in the Spanish language. In reviews of zarzuela performances, ‘decir’ (to say) was often used interchangeably with ‘cantar’ (to sing), with reviewers frequently praising zarzuela singers for saying their role very well.³³ This focus on text can be connected to the concept of naturalidad or naturalismo (naturalness, spontaneity), which reviewers often invoked as a desirable quality in zarzuela performance, usually in opposition to what was heard as artificiality and affectation in other genres, such as opera.³⁴ The requirement to convey text expressively in singing was also linked to the expectation that zarzuela performers should be convincing when reciting and acting during the sections of spoken dialogue (which, as will be discussed later, were not included in recordings). Zarzuela singers were therefore expected to communicate Spanish-language text expressively and in different ways throughout a performance: to sing zarzuela well was often equated with being able to sing well in Spanish.

Within zarzuela circles, these factors contributed to defining the genre’s expressive codes and how audiences received them. Their manifestations in zarzuela were seen as being somewhat distinct from those of opera, while still sharing significant common ground. In the early decades of zarzuela development, pioneers such as Francisco Asenjo Barbieri repeatedly complained that the professors at the conservatoire, as well as others teaching privately in Spain, did not teach their students to sing in Spanish (vocal training at the time was focused mostly or solely on Italian-language materials). Musicologist Antonio Cotarelo y Mori notes that some singers, in order to be able to sing zarzuela, had to retrain after completing their degrees at the conservatoire, although he did not explain where or with whom they did this.³⁵ Composer Emilio Arrieta, who directed the Escuela Nacional de Música y Declamación from 1868, was credited with providing increased performance opportunities for students of singing,³⁶ but advocates of Spanish opera still claimed that students were not learning to sing properly in Spanish.³⁷ This further suggests that zarzuela composers and supporters acknowledged that singing in Spanish necessitated approaches and techniques that were at least in some respects different from those required to sing Italian opera, even though details of such approaches and techniques were not explicitly explained.

It is the recordings themselves which can therefore offer some starting points in terms of understanding how the principle of communicating a text expressively might have been realized in performance. In fact, even a cursory listen to a few recordings quickly reveals that the ability to sing in Spanish did not necessarily equate with text intelligibility. While this might have to do with technical limitations (especially in the case of earlier recordings), what the body of recordings suggests overall is that intelligibility was not a gold standard and that a performance could preserve some of the text’s expressiveness even if not every word was clearly understandable.

One such example is Pilar Duamirg’s recording of the well-known aria ‘Pensar en él’ from Marina, recorded for Parlophon in 1930.³⁸ Duamirg’s pronunciation is almost unintelligible; however, she conveys a strong sense of spoken speech. She subtly manipulates note values, shortening some and lengthening others, in order to adapt her singing to the cadences of spoken Spanish. Duamirg introduces localized rubato in key places to achieve the same effect. Central to this study, she employs portamento to emphasize key moments in the text and to highlight some of her tempo modifications. This performance, while musically interesting on its own, is better understood within the context of several decades of changing conventions regarding portamento and how it was documented in recordings, the focus of this Element.

Zarzuela on Wax Cylinders, 1896–1905

The first recordings ever made of zarzuela are likely lost forever. These took place during public demonstrations of the newly invented phonograph that took place in Spain between 1878 and 1882 and then more frequently throughout the early to mid 1890s.³⁹ At that time, sound recording technologies had not yet developed into machines intended for domestic use. Instead, Spaniards would gather in theatres, cafes, clubs, churches or larger private homes – or visit commercial premises (salones fonográficos) – to watch an operator demonstrate the phonograph. The main purpose of these demonstration sessions was for audiences to be reassured first hand that the phonograph was indeed capable of recording and playing back sound. It is likely that the resulting recordings were discarded quickly after the demonstration, with little if any thought given to their preservation.

To show off the phonograph’s capabilities, the operator would normally choose a range of sounds with which audiences would be familiar, such as military calls, animal sounds, jokes and short speeches, opera arias, traditional songs and, indeed, zarzuela romanzas. On some occasions, some of these would be recorded in front of the audience by local singers or speakers and then immediately played back, a process which reassured those present that the operator had not manipulated or faked the recording in any way.⁴⁰ Press reviews typically speak of the audience’s wonderment at the phonograph’s capabilities, but they tend to provide little or no detail on the titles of the zarzuela numbers that were recorded, who sang them and whether the singers changed their performance in any way when recording.

There are a few exceptions, such as the press notices for a private salón fonográfico that a Mr Pertierra kept open in Madrid in 1894 and 1895. From advertisements, we know that Pertierra’s zarzuela repertoire included selections from recent works such as El gorro frigio (1893), Certamen nacional (1888) and La verbena de la Paloma (1894) sung by well-known performers active on the Madrid stages, namely Emilio Mesejo, Joaquina Pino, Loreto Prado and Lucrecia Arana.⁴¹ It is remarkable that Pertierra managed to secure some of the most successful and best-known performers of the time, since, as will be discussed later, in the early years of commercial recording it was more common for companies to employ rank and file, lesser-known singers. Mesejo, in particular, never recorded in the era of commercial phonography and Arana only did so in a sustained way during the final years of her career.⁴² The lack of surviving pre-commercial cylinders means that we might be missing out on potentially important sources that would allow us to obtain a fuller picture about zarzuela performance practices in the 1890s (and even earlier), since the careers of some of these singers went back to the 1870s or even the 1860s.

The official arrival of commercial phonography in Spain can be dated to December 1896, with the opening in Madrid of the first gabinete fonográfico, Hugens y Acosta. Gabinetes were small establishments – often operating as a sideline to an established business, such as a pharmacy or an optician’s office – that imported phonographs to sell to the public. They also recorded their own wax cylinders, using Edison machines, with local singers and instrumentalists. The gabinetes era was short-lived. Hugens y Acosta, which was also the last of them to close, did so in 1905. They were unable to face competition from the international firm of Gramophone Company, Ltd., which Emile Berliner founded in London in 1898. During these nine years, about forty gabinetes operated for various lengths of time in several cities and towns across Spain, but most were in Madrid, Barcelona and Valencia.⁴³

The number of surviving cylinders from the gabinetes currently held in public collections now amounts to about 1300 and about 300 of these are of zarzuelas. While this seemingly small number appears to constitute only a small part of the total output of the gabinetes, recording zarzuela actually constituted a sizeable part of the activities of these firms. Zarzuela was actually the third most represented genre among these cylinders, after brass band music – recordings of which might have been so plentiful not necessarily because of the genre’s popularity but because it was technologically easier to record than other genres – and opera.⁴⁴ There is also evidence that zarzuela was being recorded outside Spain in these early years, notably by Pathé Frères in France. Pathé cylinders were not playable on Edison machines and vice versa, so the circulation of Pathé recordings and devices might have been somewhat limited during these years in Spain. Only one Spanish collection of Pathé cylinders containing zarzuela examples is known to survive.⁴⁵

The reasons why zarzuela was one of the first genres to be extensively recorded in Spain are easy to understand. The genre was extremely popular with audiences, especially in the cities, most notably Madrid. The subgenre known as género chico in particular had evolved into a veritable mass entertainment industry. To understand the popularity of género chico, we need to go back to the birth of zarzuela in its modern form in the mid nineteenth century.⁴⁶

At this time, the predominant zarzuela format was zarzuela grande, as introduced by Francisco Asenjo Barbieri in Jugar con fuego (1851). It consisted of three acts and featured a range of musical numbers linked by spoken dialogue. Zarzuela grande drew upon European models such as Italian bel canto and French opéra comique and infused these approaches with Spanish traditional and historical dance rhythms as well as with Spanish settings and stories to create something distinctively Spanish. Zarzuela grande was at first remarkably popular with middle class audiences in Madrid and elsewhere, but from the mid 1860s, new works were failing to gain the accolades of their predecessors, a situation that put the future of the genre at risk.⁴⁷

A lasting solution to the crisis came through the teatro por horas format. Madrid audiences had been becoming increasingly enthralled with teatro por horas (literally, theatre by the hour) since 1868,⁴⁸ when theatre impresarios replaced their customary full length spoken plays, which typically lasted three to four hours, with four one-hour plays. Performances started in the late evening and continued until midnight.⁴⁹ Tickets for each play were sold separately, which increased profits for impresarios and minimized risk since unsuccessful or unpopular plays could be easily replaced. While teatro por horas was initially used only for spoken plays, zarzuela producers also eventually adopted the format. The shorter, one-act zarzuelas started to be collectively known as género chico (literally, the small genre), with Federico Chueca’s La canción de la Lola (1880) generally being acknowledged as the first work in the new format.

Apart from being shorter to fit the parameters of teatro por horas, género chico works were different from zarzuela grande in ways that could sometimes demand different approaches to performance. The subject matter of género chico works was comic and lighter than those of zarzuela grande; plots and characters were subjected to a higher degree of standardization, as suited the commercial nature of the new subgenre. Furthermore, the music started to show greater influences from Spanish traditional and popular genres (as opposed to Italian bel canto) and strophic songs became more frequent. The latter facilitated memorization in a context where actors were forced to learn new material quickly, since the same music was sung to different words. Like zarzuela grande, género chico still used predominantly Spanish settings and themes, but now the pueblo was placed firmly at its centre. This meant that it focused on the Spanish working and lower middle classes, including peasants, skilled and unskilled urban workers, artisans, small business owners such as shopkeepers and, in some cases, lowlife characters. The genre’s portrayal of the pueblo was designed to be palatable and appealing to the bourgeoisie, the upper classes and the aristocracy,⁵⁰ while at the same time allowing the working and lower middle classes to recognize themselves in the characters.⁵¹ As a result, the class dimension of zarzuela audiences expanded with respect to what had been the norm earlier.

As género chico developed, longer-format zarzuelas were still being performed and written (e.g., Ruperto Chapí’s La tempestad (1882) and El rey que rabió (1891), both of which were immensely popular), although their appeal remained more limited to the middle and upper classes.⁵² Even though in the late nineteenth century separate companies and theatres existed for género chico, zarzuela grande and opera⁵³ and zarzuela grande performers were expected to have received operatic training in ways that género chico performers were not, an examination of the career patterns of individual singers shows that some noted género chico performers, such as Lucrecia Arana, did indeed have operatic training⁵⁴ and that, all in all, the genres were relatively permeable. Zarzuela singers moved not only between zarzuela and opera but also between zarzuela grande and género chico as well as between zarzuela, lighter forms of entertainment and spoken theatre. Decisions to move between genres depended on one’s vocal and acting ability (which changed with age) and also on the perceived prestige of different styles, financial considerations, available opportunities and personal preference.⁵⁵

It was against this backdrop that the gabinetes fonográficos were first established and started recording zarzuela as a vibrant and lively stage genre. The gabinetes had to contend with a number of limitations and challenges (technological, commercial, cultural and sociological) that considerably shaped what they could offer and what we can hear in their recordings nowadays. Some of these limitations, challenges and particularities are common to the global recording industry at this point in time, while others are more specific to Spain. First, the gabinetes did not put dates on their cylinders and catalogues were scarce, so attempting to draw firm conclusions regarding how styles might have evolved between the last years of the nineteenth century and the early years of the twentieth is problematic.⁵⁶ Second, recordings made by the gabinetes fonográficos were almost always unique. Due to technological limitations, duplication was generally only possible by means of a pantograph, which meant that resulting copies were limited in number and of poor quality. Many of the most active Spanish gabinetes took pride in the fact that their products were one-of-a-kind and individually crafted, rather than mass-produced.⁵⁷ For matters of performance practice, this has significant implications in that gabinetes and singers were likely more concerned with ensuring that each of their recordings was good enough to be sold rather than with producing a perfect version of any given piece. Importantly, these recordings therefore preserve some of the spontaneous, ‘in the moment’ quality of live performance.⁵⁸

Additionally, the gabinetes faced the challenge of making their customers accept the new medium of recording as a faithful representation of reality. This was a challenge that Spanish recordists shared with others elsewhere in the world, but the situation in Spain had some particular characteristics. Familiarity was key: gabinetes did not try to introduce their audiences to new repertoires, but rather focused on music which buyers would know well, be it older, established works or the latest hits. But the genre’s ever-evolving, multimedia practices were not easy to capture on record, particularly on two-minute cylinders.⁵⁹ Zarzuela, particularly género chico, was a robust industry in which the act of seeing any stage performance would have been significantly coloured by the experience of seeing similar works live on stage on a regular basis, perhaps even daily.⁶⁰ The gabinetes certainly made some attempts to capitalize on live performances. A majority of the gabinetes in Madrid, for example, were located near the género chico theatres, where they could have attracted theatre-goers as customers and also have ready access to singers.⁶¹

The gabinetes were also forced to make adaptations that quickly became widely accepted conventions in the nascent recording industry. Two of these are of particular importance. First, the limited duration of cylinders made it impossible to seamlessly record a full zarzuela and even individual numbers had sometimes had to be shortened through cuts (rather than through faster speeds).⁶² Second, spoken dialogue was not recorded; only the musical numbers were.

Even though it would have been possible for gabinetes to split a long number across two (or more) cylinders, examples of this practice are very rare, which suggests that the illusion of seamless performance was, from early on, important to early recordists and listeners. There is no evidence, moreover, that recordists were at this stage thinking beyond individual numbers: gabinetes were not producing ‘full’ zarzuela or opera recordings by selling several numbers of the same work together (perhaps, at least in part, because of the high costs this would have incurred) and similarly singers did not have the opportunity at this stage to construct a full-length performance on record (as they would later on).

Although spoken dialogue was regarded as an important part of the genre in zarzuela circles (with performers being expected to recite dialogue convincingly), dialogue never made it onto zarzuela recordings, other than quick interjections in the middle of a number. These typically occurred in zarzuelas set in urban, working class environments such as Agua, azucarillos y aguardiente and La verbena de la Paloma. The question as to why dialogue was systematically cut does not have a straightforward answer. The gabinetes also sold spoken word recordings that featured comedy routines, short stories or speeches (as did other manufacturers of cylinders). These were sold at lower prices than music cylinders and were eventually discontinued, which implies that they were not very popular with audiences. However, in the same way that the gabinetes experimented with spoken word cylinders, they could also have experimented with zarzuela dialogue, though there is no evidence to suggest that they did so. In contemporary opera recordings from around the world, recitative was also commonly left out, so it is plausible that gabinetes simply took their cue from elsewhere and adopted the practice to their own conventions.

Despite the gabinetes’ significant technological and cultural breakthroughs, their impact on broader society remained limited at the time and it should not be assumed that music listening practices were transformed in any meaningful way. In particular, the zarzuela theatre industry remained relatively impervious to new technological developments, with the obvious exceptions of the singers who recorded them. Theatre and zarzuela publications of the time, such as the magazines Juan Rana and El arte del teatro, only mention the phonograph and mostly under the guise of general commentary or satire rather than offering considered commentary of individual recordings. Composers at the time concerned themselves more with protecting their intellectual rights in order to secure financial benefits from live performances than engaging with the recording industry. Since the gabinetes were only a small industry, it seems likely that composers did not see many financial advantages in working with them.

These considerations notwithstanding, a significant number of the surviving wax cylinders reflect high levels of vocal technique and artistic quality. One example is Eloísa López-Marán’s recording of ‘Brindis’ from Manuel Fernández Caballero’s popular zarzuela La viejecita (1897).⁶³ This performance illustrates how portamento might have been used at the time, in combination with other performing decisions, to create an expressive performance that attempted to somehow mimic the shape of the text and convey an impression of spoken speech.

A dozen cylinders by López-Marán survive, making her one of the early zarzuela performers for whom we have the most recordings. This quantity suggests that she recorded extensively, even though her stage career was not particularly illustrious. López-Marán studied with the then-famed teacher Napoleón Verger, but notices about her stage career are limited to a handful of reviews for performances in Madrid opera and zarzuela theatres between 1894 and 1901.⁶⁴ She was typical of most singers who recorded for the gabinetes in that she was not a star performer. (This is consistent with practices elsewhere.) The recording industry at this point was considered too much of an uncertainty to attract sustained interest from major singers. Surviving recordings as well as catalogues from gabinetes suggest that companies might have relied to a considerable extent on a relatively small number of vocalists who recorded prolifically⁶⁵ and who likely stood out for their ability to quickly adapt to recording technologies and the recording studio.

López-Marán, in her recording of ‘Brindis’, suggests that she was one of these singers (see Figure 1). At the start of the passage, the tempo moves steadily (bars 1–7), and the portamentos are discreet (in bars 3 and 5), more akin to a well-sustained legato that López-Marán might have used to efficiently prolong the vocal line than to achieve a deliberate vocal effect. The next portamento (bar 7) is obviously more deliberate: clear and well-controlled, it moves from the fermata down to the leading tone, in what amounts to a very commonplace use of portamento in recordings of zarzuela from these and later days. The dotted quavers soon thereafter might have invited further discreet portamentos (bar 11), but López-Marán chose not to do so and therefore isolated the word ‘ingratos’ (ungrateful) from the rest of the line in an obvious dramatic effect. A few bars later, the portamentos are again prominent (bars 14–15), along with small-scale decelerations of the tempo. When combined, these performance choices provide a sense of closure to the section while building up expectations for the next one. We could have expected a portamento on the word ‘ellos’ (bar 15), as portamento was also exceedingly common in descending fourths, but López-Marán again decided otherwise. At the same time she chose not to slow down any further, perhaps to avoid a sense of exaggerated deceleration.

Figure 1 Bars 5 to 20 from ‘Brindis’, from Manuel Fernández Caballero’s La viejecita. Portamento and fermatas have been added in grey to reflect Eloísa López-Marán’s performance

López-Marán’s performance of this passage reveals some underlying principles concerning the use of portamento that recur in several of these earlier (and also later) recordings. Portamento is often combined with tempo decelerations so that the singer can emphasize not so much a specific word but rather a part of a sentence, thereby giving dramatic shape to the line. Remarkably, La viejecita is a género chico work, which in principle might have been thought of as less receptive to elements of operatic performance like portamento. López-Marán’s recording, however, confirms that portamento could indeed be used liberally and intelligently within the subgenre.

There are a limited number of zarzuela romanzas for which more than one cylinder survives (typically no more than three or four). These are generally from works that enjoyed successful premieres during the time of the gabinetes. Such is the case with both La viejecita and Fernández Caballero’s other hit, Gigantes y cabezudos (1898). The plot of Gigantes y cabezudos referred to the then-current Spanish-American war over Cuba. The zarzuela became a success in large part because of how it depicted the ambivalent attitude of many Spaniards towards the war, which was largely seen as a heroic yet vain effort that came at an enormous cost to Spanish society. Comparing these recordings can provide us with a glimpse into the general parameters that governed the use of portamento and how a performer might have chosen to use portamento in a specific performance following these guidelines. At the same time they preserve some of the spontaneity and live quality of recordings at the time. Illustrative examples are cylinders of ‘Esta es su carta’ from Gigantes y cabezudos as recorded by Adela Taberner⁶⁶ and by Marina Gurina⁶⁷ (see Figures 2a and 2b) and ‘Canción del espejo’ from La viejecita by Blanca del Carmen⁶⁸ and by Pepita Alcácer⁶⁹ (see Figures 3a and 3b).

Figure 2a and 2b Bars 40–51 from the Tiempo de vals lento ‘Esta es su carta’, from Manuel Fernández Caballero’s Gigantes y cabezudos, with portamento added as performed by Adela Taberner and Marina Gurina respectively

Figure 3a and 3b Bars 4–11 from ‘Cancion del espejo’, from Manuel Fernández Caballero’s La viejecita, with portamento added as performed by Blanca del Carmen and Pepita Alcácer

Such recordings allow us to observe some of the tacit rules that governed the use of portamento. First, descending portamentos were more common than ascending ones, although the latter definitely existed. The appearance of ascending portamentos actually contradicts one of the very few specific instructions about portamento found in singing treatises such as El cant (llibre per al cantat i per l’aficionat), written several decades later by the celebrated tenor Emili Vendrell, who said that ascending portamento was ‘anti-artistic and in bad taste’.⁷⁰ Second, portamento tended to be used for intervals of a fourth or larger, rather than for smaller ones. For smaller intervals, when portamento does appear, it tends to take the form of an exaggerated legato, as in the López-Marán recording (see Figure 1). This approach becomes more prominent in later recordings, as we will see. Third, portamento tended to appear more at the ends of phrases than at the beginnings.

The approaches to musicianship displayed in the recordings by López-Marán, Del Carmen, Taberner, Gurina and Alcácer at first might sound unusual because of the portamentos and tempo fluctuations, which become increasingly apparent after several sessions of attentive, unprejudiced listening. Some performers used portamento more profusely and others more sparingly, but in every case it eventually becomes possible to recognize the expressive intentions achieved through portamento.

Other gabinete recordings, however, might be more difficult to interpret in expressive and musical terms. This comes as a useful reminder of two important considerations: 1) the need to disentangle what we hear in early recordings in terms of performance practice and 2) the multifaceted, hybrid nature of zarzuela. Both aspects are central to this Element and therefore deserve further exploration.

To illuminate the first point, I will focus on a recording of ‘Salve’ from Gigantes y cabezudos, the final number of the zarzuela, performed by an unnamed chorus at the Corrons gabinete in Barcelona. Choral recordings are relatively rare within the earlier zarzuela recording output (as is typical of the earlier phonograph era). The chorus often appears at pivotal moments in many zarzuelas, including Gigantes y cabezudos. Here, it represents the people of the pueblo, who either come together and sing of their identity or express patriotic feelings, the latter more often than not with a hint of humor and irony. Recording even a small group of singers multiplied the difficulties of recording a single soloist.⁷¹ Some gabinetes and private recordists eventually developed acceptable solutions to record zarzuela choirs in what was likely a highly experimental process,⁷² but the number of such recordings remained small.⁷³

In the ‘Salve’ recording, portamento is employed on nearly every descending fourth and fifth interval; this is particularly evident at the end of phrases. By contrast, when considering the small number of other choral recordings from the time, portamento is either used sparingly or not at all.⁷⁴ When recording a soloist, the decision to use or not use portamento could have been made on the spot. It is less likely that this was the case when it came to ensembles, which would require some rehearsal in their collective use of portamento. In the portamento-rich ‘Salve’, the use of portamento is so pronounced that the performance risks becoming unmusical. This suggests that the use of portamento was a very deliberate decision and that the choice was perhaps not so much musical as it was technical. Portamento, indeed, could have functioned here as a kind of extended legato that helped the choir keep a semblance of vocal line while singing together and was something with which they experimented in the recording room in order to alleviate the problems inherent in recording even small choirs.

Such a hypothesis opens up the possibility that performers – especially those who were more experienced making recordings or those who had worked under the direction of a knowledgeable recordist – consciously adapted their performing styles to suit the phonograph. With their awareness that some dimensions of the sound could not and would not be captured, they instead tried to compensate or replace those aspects with others. I have suggested elsewhere that the briskness of the tempo in some of these earlier recordings might have been another result of these conditions⁷⁵ and similar conclusions concerning such compensatory mechanisms have been advanced in recent empirical studies involving modern performers recording on early-style phonographs.⁷⁶

Another example of an outlier recording is that of ‘Canción de Paloma’ from Francisco Asenjo Barbieri’s classic zarzuela El barberillo de Lavapiés (1874) by an unnamed singer for Pathé, the budding French multinational that also occasionally recorded zarzuela in these years. This recording, therefore, offers a rare example of an early zarzuela recording made outside Spain.⁷⁷ As such, it provides a broader context for the practices developed by early Spanish recordists and singers in manipulating and interacting with technology. The execution of portamento in this recording might seem disappointing, even when considering the limitations of technology. Sometimes initial and/or final notes are noticeably out of tune, which could reveal a lack of technical control on the part of the singer. We cannot know for sure, however, whether these types of portamentos would have been heard as acceptable by zarzuela audiences, but we can contextualize the singer’s portamento technique alongside various other performance aspects of the recording. Elsewhere in the romanza, the acciaccaturas are not particularly well executed and musical accents are often misplaced. It is entirely plausible that this recording was not particularly representative of what a good performance would have sounded like on stage. We know that wax cylinder recordings often featured lesser-known singers, some of whom may have been mediocre or inexperienced. We also know that some singers became anxious and struggled when first going into the recording room.⁷⁸ Nervousness and an inability to keep a steady airflow might well be the root of a number of the issues described above. Singers on Pathé zarzuela recordings are often unnamed, which reveals a certain lack of care in recording the genre. This contrasts greatly with the close attention to detail among some of the gabinetes.

The second consideration, the hybrid nature of zarzuela, is a key aspect of the genre that in recent years has gained increasing importance in its historiography. Acknowledging zarzuela’s hybridity has allowed scholars to reaffirm and strengthen its connections with similar vernacular genres around Europe⁷⁹ and this assertion contrasts greatly with earlier writings that insisted on the exceptionality of the genre and which typically dismissed influences from non-Spanish genres, particularly from Italian opera, as contamination.⁸⁰ Zarzuela, despite being a typically Spanish genre featuring numbers inspired by traditional forms such as jota and guajira, was by no means isolated from musical developments outside Spain. Influences at different points in the history of zarzuela included Italian bel canto, French opéra comique and opérette and German-language operetta, as well as cabaret and other forms of musical theatre, as discussed above.

Early recordings prove that hybrid approaches influenced zarzuela performance and the use of portamento. For example, in a recording of ‘Si las mujeres mandasen’ (also from Gigantes y cabezudos) made by a señorita Martínez for the Barcelona gabinete Corrons,⁸¹ the singer does not use any portamento. This makes her recorded performance highly distinctive from those of Gurina and Taberner in the same role. Martínez’s diction comes across as remarkably clear, though her voice does not seem to have the resonance of Taberner’s and especially Gurina’s.

Martínez’s style seems more akin to the popular and traditional styles of singing that focused on clear text enunciation at the expense of vibrato and uniform timbre.⁸² In these styles of singing, executing a well-controlled portamento will have likely been more difficult. Moreover, it might have been also the case that using portamento in such contexts would have been seen as an affectation that did not align well with the rest of the aesthetic.

Martínez’s performance on this recording makes more sense when it is placed within the history of zarzuela in the decades preceding the arrival of the phonograph in Spain, as well as within what we know about the singing profession during these years. Although género chico preserved the Italian and French influences that had informed earlier zarzuela, it also showed an increased capacity to absorb other musical influences and performance genres (e.g., dance, drama).⁸³ While some singers such as Lucrecia Arana did have operatic training, others simply learned on the job.⁸⁴ Press reviews suggest that these performers were stronger actors than singers.⁸⁵ Martínez’s recording is likely a representative example of this actor-driven approach to performance.

In contrast to Martínez’s approach, those of Gurina and López-Marán can likewise be understood within the context of the shifts in vocal technique that were taking place during the late nineteenth century. This was also the time when the idea of vocal registration – the uniformity of timbre across the different registers of the voice – started to impose itself on the operatic world, especially as a requirement for singing verista roles like those found in the operas of Mascagni and Leoncavallo. This uniformity of timbre was achieved through the raising of the soft palate and the lowering of the larynx, which elongated the vocal tract to maximize resonance and vibrato.⁸⁶ Portamento might have been easier to execute for singers familiar with this technique, since register uniformity would have made portamento sound smoother. The vocal tract modifications required to achieve uniformity in register, however, often came at the cost of clear diction.

Hybridity in performing styles, however, did not simply mean that different singers with different vocal techniques and approaches could perform the same roles, as was the case with Pilar in Gigantes y cabezudos. It also meant that performers could adapt their own styles to a certain extent by incorporating certain elements (or not), depending on the context. This was the case, for example, with vibrato. While the new operatic vocal techniques discussed above implied that vibrato had to be continuous, in zarzuela recordings of this era (and beyond), this is not the case, for it is heard only on certain notes or during specific passages.⁸⁷ The same is true of portamento. Choosing whether or not to effect a portamento could mark a certain romanza as being more popular than operatic in nature.

For example, in Emilio Arrieta’s Marina (first a zarzuela grande, then an opera and consistently a popular choice among gabinetes),⁸⁸ the soprano and tenor romanzas routinely employ portamento,⁸⁹ while baritone Roque’s seguidillas do not.⁹⁰ The seguidillas offer one example of the sort of numbers inspired by traditional and popular musical forms that proliferated in zarzuela, mostly in género chico but also in larger scale works such as Marina. In their recordings of seguidillas, singers generally limited the use of the expressive devices frequently found in zarzuela recordings that were more akin to opera (e.g., ornaments, tempo changes and vibrato) and instead chose fast tempos and vigorous deliveries that seem to show little influence from the more operatic approaches to singing.

In conclusion, while the number of zarzuela cylinders that survive is comparatively small and assuming that most of them would have been one-offs, due to the difficulties in the duplication process, the picture that emerges from these scattered examples is as follows. Different approaches to zarzuela performance coexisted under the general principle of expressive delivery of text. Portamento often, but not always, became an audible marker that differentiated the operatic approaches from the more popular and vernacular ones.

Zarzuela on 78 rpm Discs, 1899–1924

Between 1899 and 1905, the gabinetes coexisted with a newer form of technology that would eventually become dominant and drive them out of business: the gramophone. A key difference between the technologies was that, unlike cylinders, gramophone discs could be duplicated easily and reliably. Gramophone, the London-based manufacturer that was so central to the technology that it took the company’s name, developed a strategy that did not need to remain localized, as was the case with the gabinetes. Hence, the entire recording industry quickly became more fast-paced and transnational than ever before.

The process was revolutionary. First, Gramophone dispatched scouts around the world. Second, they then either opened branches or absorbed smaller, local labels to establish their foothold in various international markets.⁹¹ These branches allowed Gramophone to record vernacular repertoires while developing various local markets for their products. Czech-language recordings made in Europe, for example, could be sold both in their country of origin and among Czech émigré communities in the USA.⁹² Furthermore, a small number of such recordings of vernacular and indigenous repertoires could be marketed to an international audience with an appetite for exotic recordings.⁹³

It was during these years that an increased circulation of recordings, including ones of zarzuela, began. During the first quarter of the twentieth century, this mobility would result in a more globally cohesive set of musical tastes and listening habits than had previously been the case.⁹⁴

In Spain, this process can be dated back to 1899, the year that Gramophone scouts first visited Madrid and Barcelona. They returned to the same two cities in 1900 and 1902 and in 1903, Gramophone opened a branch in Barcelona. At the time, the gabinetes were still operating and many sold Gramophone products alongside their own. It did not take Gramophone long, though, to force the gabinetes out of business, with the last one, Hugens y Acosta, closing in 1905, as mentioned in the previous section. Despite rivalries between Gramophone and the gabinetes,⁹⁵ significant overlaps existed between the two technologies. It is telling, for example, that during their initial visits to Spain, Gramophone signed a significant number of singers who had already recorded for the gabinetes. The firm may have approached singers such as Bernardino Blanquer, Ascensión Miralles, Daniel Blanquells, Josefina Huguet, Blanca del Carmen and Manuel Figueroa because of their familiarity with the mechanics of a recording session at a time when few performers understood such things. Or, it is quite possible that these singers might have put themselves forward for the same reason.

Since gramophone discs were still limited in length, recordists and singers did not look beyond individual numbers, following the practice of the gabinetes. Many earlier gramophone discs and gabinete cylinders, therefore, can be thought of and understood within the same contexts. For example, the internationally successful baritone Marino Aineto recorded the character Roque’s ‘Seguidillas’ from Marina for Gramophone in 1902. This was around the same time as the two gabinetes recordings discussed in the previous section. While Aineto’s voice is obviously attuned to operatic modes of delivery, the three performances can certainly be understood as reflecting the same underlying style and approach to expressiveness with their brisk, regular tempos and sharp diction.⁹⁶

There were, however, key distinctions between gramophone and gabinete recordings. First were the differences in technologies and what kinds of sounds recorded better in each format. Second, since Gramophone had access to some of the most prominent opera singers in the world through its international branches, it had little incentive in Spain to record local opera stars and instead focused on zarzuela.⁹⁷ Third, Gramophone’s recording and marketing decisions, by virtue of its global reach, were more complex than had ever been the case with the gabinetes.

Concerning technology, gramophones were said to lack the sensitivity and artisanal nature of phonograph recordings, a fact that the gabinetes fonográficos often used as the main criticism against their competitor.⁹⁸ On the other hand, this meant gramophones were not as prone to accidental, unwanted moments in respect to live sound, such as pitch distortions. Even though the gramophones were still lacking in many ways, they were able to capture a broader range of sounds in an acceptable way than the earlier technologies.⁹⁹ For example, since recording string instruments remained difficult, gramophone recordists were able to develop effective ways of recording singers accompanied by small brass orchestras. This is the kind of accompaniment – as opposed to piano – that we hear in most zarzuela recordings from the period. Gramophone discs could also capture dynamics and timbres in more realistic ways, which opens up the possibility that some of the compensation mechanisms discussed in the previous section when it came to recording zarzuela numbers might have evolved or even disappeared. In the context of these slightly later disc recordings, the choral recording of ‘Salve’ from Gigantes y cabezudos discussed in the previous section stands out as an anomaly, which suggests that such widespread use of portamento was not the norm and might have been used to address a specific technical-technological problem, namely overemphasizing one expressive device which recorded well (i.e., portamento) to compensate for ones that did not.

Since zarzuela could not be easily recorded anywhere else but in Spain (with some isolated exceptions in Latin and North America), it was on this repertory that the Spanish branch of Gramophone focused.¹⁰⁰ Other companies followed similar strategies, such as Odeon, which was established in 1903 as a label of the International Talking Machine Company, then the leading rival of Gramophone and in January 1904 became part of the Carl Lindström Company. Odeon and Gramophone became the leaders in the zarzuela recording market and remained so during the following decades. At the same time, recordings were starting to become a much more attractive proposition for singers, particularly after Caruso’s recordings for Victor in 1904 proved that making recordings could be a lucrative career move for a performer with an already established stage reputation.

The mass-production and mass-distribution processes of multinationals meant that recordings by international celebrities could now be sold everywhere. This reduced the need for local companies to produce their own recordings. For some time after 1904, Gramophone kept hiring some of the singers it had employed in its earlier years, such as Rafael Bezares (a singer with a very prolific recording career), but the company gradually changed its hiring practices to fall in line with the rest of the industry. As a result, recording zarzuela became increasingly concentrated in the hands of a much smaller number of performers than before. These were now normally singers who had achieved success on stage, many of whom were part of a new generation of performers who began their careers after the era of the gabinetes. They included the husband and wife team of baritone Emilio Sagi Barba and tiple Luisa Vela,¹⁰¹ baritone Inocencio Navarro and tenor José Palet. These singers form the first generation of zarzuela singers who from the start consistently made recordings integral to their professional careers. In this new era of recording technology, we can assume that particular choices in terms of performance practice do not necessarily mean a lack of ability on the part of the singer, but rather are conscious decisions.

Apart from changes in recording technology and the recording industry itself, it is likely that the dynamics within the zarzuela sphere also contributed to these developments. Around 1900, género chico started to decline in popularity, which led to the rise of the so-called géneros frívolos. These were comparable to non-narrative genres from outside Spain such as music hall, cabaret and vaudeville. The most important subgenres of géneros frívolos were cuplé and variedade. A cuplé consisted of a single song, normally sung by a female performer, that was often erotic in nature and therefore capitalized on the singer’s sexualized stage presence. Variedades offered a series of singing, dancing and comedy numbers, usually performed by different artists, that were sometimes tied very loosely to a common topic.¹⁰²

Alongside the flourishing of the géneros frívolos, longer-format zarzuela was experiencing a significant revival. Thanks to influences from French, Austro-Germanic and English operetta styles, the genre experienced a resurgence with works such as Amadeo Vives’s Maruxa (1914), José Serrano’s La canción del olvido (1916) and Pablo Luna’s El asombro de Damasco (1916) and El niño judío (1918). Operettas such as Die lustige Witwe (The Merry Widow) and Die Dollarprinzessin (The Dollar Princess) were also sometimes performed by zarzuela companies in Spanish translation (as La viuda alegre and La princesa del dóla, respectively). Performers such as the aforementioned Sagi Barba, Vela and Navarro, among others, rose to fame as the champions of the renewed zarzuela grande and some, such as Vela, also appeared regularly in translated operettas.

After 1910, the zarzuela scene was, in some respects, the opposite of what it had been for the past three decades. New zarzuelas grandes were regularly appearing, with different levels of success. They were often produced by companies managed by the zarzuela composers themselves. A handful of these new zarzuelas remained in the repertoire, but audiences were not as drawn to new zarzuelas grandes as they once had been. This meant that impresarios had to program audience favourites alongside new works as a hopeful guarantee for a season’s overall commercial success. Familiar titles, it was hoped, would cover the costs for new works, which may not recoup their own expenses. Audiences continued to be drawn to past favourites such as Marina (Emilio Arrieta), La tempestad (Ruperto Chapí), El juramento (Joaquín Gaztambide), El salto del pasiego (Manuel Fernández Caballero) and Jugar con fuego (Francisco Asenjo Barbieri).¹⁰³ Zarzuela was starting to become increasingly dependent on its past repertoire, rather than on new works. This was especially evident in género chico, which by 1910 had practically become fossilized. Earlier successes such as La revoltosa, Gigantes y cabezudos and La gran vía remained popular and an almost certain guarantee of success for any zarzuela company. The subgenre was now being seen more as a window to the past of the Spanish pueblo rather than a living art form.¹⁰⁴

The changing balance between género chico and zarzuela grande had a definite impact on the working patterns of singers and companies. Even though some theatres such as the Apolo in Madrid performed only género chico, during this period it became increasingly less common for companies to specialize in a single subgenre. Singers had to adapt to the new circumstances and perform from both repertoires while also making recordings. To some extent, their recordings reached international audiences through Victor’s Blue and Black labels and, later, on His Master’s Voice (as the Gramophone Co. became known). Zarzuela, unlike flamenco, never became globally popular; hence, the number of zarzuela recordings made outside Spain remained very small and usually featured well-known numbers performed by singers from other Spanish-speaking countries, such as the US baritone (of Spanish ancestry) Emilio de Gogorza, the Mexican baritone José Torres Ovando and the Mexican soprano Soledad Goyzueta. These singers typically did not have the experience of performing zarzuela on stage, unlike their contemporaries in Spain. They typically recorded other genres besides zarzuela, most often opera and Spanish-language song. Among the recordings made in Spain, only a small percentage made it into the multinationals’ global catalogues. Those that did sold mostly in Latin America and among Spanish-speaking communities in the United States. While it cannot in any way be said that Spanish singers changed their performance styles for the international market, it is likely that the increased distribution of recordings had an impact on their availability and quality and vice versa. With recordists increasingly acquiring experience through trial and error and star singers replacing lesser-known ones, recordings produced after 1905 would have had a better chance than their predecessors to succeed in international markets. At the same time, opportunities to enter such global marketplaces might have provided an incentive for recordists and performers to make sure that what they produced was of a very high standard.

When it comes to portamento, gramophone recordings from the early 1900s to the years around the First World War largely confirm and deepen our understanding of the practices preserved on wax cylinders. The greater number of surviving recordings offer further evidence of the different expressive possibilities associated with portamento at this time. This is evident in the soprano romanza ‘Pensar en él’, from Marina, which María Darnís recorded for Gramophone in 1914.¹⁰⁵ The opening section (‘Pensar en él/esa es mi vida’) uses portamento sparsely and, interestingly, mostly on smaller intervals of a second or a third. This aligns with the rather declamatory quality of Darnís’s voice in this passage. Portamento, following the observations of Leech-Wilkinson and Potter cited above, contributes to the overall impression of spoken speech. Portamentos become more numerous in the final section, in which they appear on practically every descending interval of a third or larger. This allows Darnís’s performance to take on a more sung and less spoken quality.

The greater numbers of surviving discs as well as the audible improvements in technology provide us with greater understanding of the hybridity of zarzuela and how this hybridity was expressed through performance practice. In the previous section, I discussed how portamento tended to be absent from the recordings of singers whose mode of vocal production was closer to that associated with popular genres, with clear diction and little vibrato. Recordings from the gramophone era confirm that operatic and more popular styles were not two separate categories but rather formed a continuum. Singers were typically placed along this spectrum based upon their technical skill, their vocal habits and preferences and their ability to access a bespoke range of expressive resources.

Two zarzuelas that were being recorded with increasing frequency during these years, both of which support this assertion, are Tomás Bretón’s La verbena de la Paloma (1894) and Federico Chueca’s La gran vía (1886). Both are classic examples of género chico. Set in Madrid, they feature a range of relatable, if generally unheroic, characters from the urban pueblo. As such, they might have invited less operatic, more spoken approaches to performance. Written evidence suggests that this was the case at their premieres, with ‘singing actors’ rather than trained singers cast in some of the main roles.¹⁰⁶ But discs show that a range of approaches to these works coexisted at the time. In Pilar Perales’s 1917 recording of La verbena’s señá Rita – a stereotypical working class matron – for Gramophone, the singer’s vibrato is appreciable and so is the uniformity of her registers. Portamento is not overly prominent, but when it does appear, it is quick and well-executed. Furthermore, following convention in more substantial romanzas, it appears at the ends of phrases.¹⁰⁷ Manuel Fernández Carbonell’s 1922 recording of La Gran Vía’s ‘Vals del caballero de Gracia’, by contrast, shows a different approach to combining distinctive performance elements, including portamento, in order to create an expressive performance.¹⁰⁸ Fernández Carbonell’s vocal placement clearly features a low soft palate. His solid diction would have surely met the requirements of naturalismo. The singer’s portamentos, however, go against the expectations of Manuel García and other famed singing teachers, who thought that portamento should be executed with proper breath support and a uniformity of tone. Some of Fernández Carbonell’s portamentos sound like they are dragging,¹⁰⁹ while others appear almost like sighs.¹¹⁰ Still others hint at a very subtle drop in pitch at the end of the note.¹¹¹ It is the latter two characteristics in particular that connect Fernández Carbonell’s performance to the comments by Leech-Wilkinson and Potter about portamento and spoken language. These portamentos sit halfway between spontaneous asides that might come up in everyday speech and intentional musical artifice.

Fernández Carbonell’s performance certainly stands out for its spoken quality. Apart from the vocal production and diction, the singer shows an outstanding ability to intelligently shorten or elongate syllables in order to craft an expressive performance (see Figure 4). If we recall the 1902 anonymous recording of ‘Canción de Paloma’ discussed above, we can observe some of the same performance techniques. Fernández Carbonell’s overall rendition, however, strikes us as more musical, with its alleged shortcomings actually being used to its advantage. This might be simply because Fernández Carbonell had a greater sense of musicianship than the 1902 singer or it could be because, by this time, performers such as himself would be more experienced in studio practices and increasingly aware of how different performance elements could complement each other.

Figure 4 ‘Vals del Caballero de Gracia’ from Federico Chueca’s La Gran Vía, bars 32–40, as written (4a) and as performed (4b) by Manuel Fernández Carbonell

A further revelation coming from these pre-electrical discs that furthers our understanding of zarzuela’s hybridity is the appearance of what we might call comic portamento. Comic portamentos were normally performed by a comic or lower class character or in a comic situation. They normally occur on larger intervals (fourths, fifths) rather than on smaller ones and when compared to a standard portamento, they tend to appear more often on ascending intervals than on descending ones. Moreover, there tends to be something in the way comic portamento is performed that invites laughter: comic portamento deviates from expectations about well-executed portamentos in such an exaggerated way that it cannot be anything other than deliberate. Sometimes, for example, the second note is stressed (whereas in regular portamento, it would normally be the first), which makes the portamento sound like a sort of springboard from the lower to the higher note rather than as a subtle slide from one note to the next.¹¹² Sometimes the portamento is executed with an obvious, hyperbolic crescendo. On other occasions, comic portamentos sound like exaggerated versions of legato that in turn try to mimic certain intonations associated with lower class characters, such as slurred speech. This is the case in a recording of La Gran Vía’s ‘Jota de los ratas’ recorded by Ernesto Hervás, Rafael Alaria and Rafael Díaz for Gramophone in 1922.¹¹³

Recordings from this time also allow us to identify changes in how certain numbers were performed and how these changes might have been tied to broader debates about the cultural significance of certain zarzuelas and their places within the genre’s canon. ‘Ay de mí’, the soprano arieta from Ruperto Chapí’s El rey que rabió, was one of the most recorded romanzas during the gramophone era.¹¹⁴ The two earliest documented recordings, those by Antonietta Martínez (1899)¹¹⁵ and Emilia Colás (1902),¹¹⁶ both feature a strong spoken quality. Despite technological issues, particularly with the Colás recording, the performers demonstrate relatively intelligible diction, which suggests a low soft palate technique. Tempos in both recordings are brisk and steady and portamento is scarce. All in all, the performances are highly reminiscent of those of popular and traditional songs, such as Roque’s ‘Seguidillas’ from Marina described above. A contemporary recording by Soledad Goyzueta, made in the United States for Edison Gold Moulded Records in 1904, also features strong spoken elements (diction remains relatively clear throughout), though – by contrast – she employs a more consistent use of vibrato as well as operatic vocal technique.¹¹⁷

In the following decade, recordings by Emilia Vergerí (1910)¹¹⁸ and Luisa Vela 1915)¹¹⁹ of the same romanza reflect a large-scale shift in approach. Both singers, following Goyzueta, employ modes of vocal production that are more operatic in style, with a high soft palate and more constant vibrato. They are not identical, for Vergerí’s voice is lighter and Vela’s is more lyrical. The tempo on these recordings is considerably slower than on the two earlier ones (and also slightly slower than Goyzueta’s), which allows for greater fluctuations in tempo fluctuation and a more generous use of portamento, particularly at the ends of stanzas (see Figure 5).

Figure 5 ‘Ay de mí’ from Ruperto Chapí’s El rey que rabió, bars 27–33, with portamentos added to reflect the performances of Emilia Vergerí (5a) and Luisa Vela (5b)

The differences between the earlier and later recordings suggest that a reconceptualization of the role of Rosa in El rey que rabió might have been happening on Spanish stages at this time, one that reflected not only the hybridity of the genre but also shifts in its fundamental aesthetics. El rey que rabió is a full length zarzuela, with significant comic and even grotesque elements, alongside an important presence of the pueblo. The plot follows a young king (a trouser role) who is intent on finding out how his subjects really live by mixing with them, incognito, in a range of situations. The approach followed by Colás and Martínez in the earlier recordings would have served to emphasize Rosa’s humble origins, making her more akin to género chico female characters in her simplicity and lack of artifice. Vela’s and Vergerí’s interpretations, on the other hand, would have emphasized her status as a dramatic heroine, someone who has to overcome a number of obstacles to finally marry the king. This gives a sense of gravitas to the character at a time when full-length zarzuelas, both new and old, were returning to fashion.

Electrical Recording Takes Over, 1925–1936

The introduction of electrical recording in the mid 1920s ignited a seismic paradigm shift in the recording industry. The ways in which the sounds of voices and instruments were captured improved significantly and listening to recordings became a more aesthetically satisfying experience.¹²⁰ What one could listen to in the comfort of one’s own home was not just closer to reality in terms of fidelity than it had been previously, but it was experienced in a way that was more real than reality, for listeners could now appreciate nuances that might have gone unnoticed in the concert hall or theatre.¹²¹ These technical improvements, therefore, began to redefine listening practices worldwide. Prolonged, solitary listening for pleasure became a much more compelling experience than it had been just a few years earlier. Hence, the persona of the discophile (or record listener) – a keen listener equally interested in the technological and artistic aspects of recorded sound – started to emerge.¹²²

In many respects, the technical differences between pre-electrical and electrical recordings favour the latter, even when filtered through the prism of digitization. We can hypothesize that these technological advances must have been even more striking to 1920s listeners, recordists and singers, especially since they resulted in quick, immediately noticeable changes in how zarzuela recordings were being made. Voices could now be captured more accurately, which made for more realistic and pleasurable listening experiences without the shrillness that was sometimes present in the past when recording higher voices. The new technology was also much better for recording string instruments and so there was a shift from the brass arrangements used to accompany singers on cylinder recordings back to the original string-dominated orchestrations. Dynamics were also given a new lease of life, as it was now possible to record with greater accuracy the filados (e.g., a crescendo followed by a decrescendo, as in the Italian messa di voce) that some of the most technically competent singers added to their use of portamento. Prime examples of this technique can be heard in recordings by the baritone Marcos Redondo, one of the most celebrated zarzuela performers of the time and also a prolific recording artist.¹²³ Recordings by Redondo and his contemporaries suggest that the relationship between recorded and live music was again being renegotiated. Technological advancements would have allowed singers to take their stage mannerisms into the recording studio without needing to compensate as much for the loss of certain aspects of live performance, such as dynamics. For example, in recordings from the early electrical era, tempo modifications occur less frequently than in previous recordings and when they do appear, the rate of acceleration or deceleration tends to be less.

In other respects, the impact of electrical recordings on performance styles and the listening experience was more gradual and broadly based. Recordings were intertwined with a range of musical, cultural and social developments that transpired over years or even decades. The early years of electrical recording took place during a particularly fruitful time for zarzuela grande. Amadeu Vives, after some successes in the second decade of the century, went on to write what is perhaps his best-known work today, Doña Francisquita, in 1923. Other noted composers were also establishing themselves in the newly thriving genre, including the team of Reveriano Soutullo and Juan Vert (La del soto del parral (1927)) and Jacinto Guerrero (Los gavilanes (1924) and El huésped del sevillano (1926)). The first half of the 1930s (more specifically, until the beginning of the Spanish Civil War in 1936) saw the rise of two composers who came to dominate Spanish theatrical life at the time and who can be rightly described as the two last great zarzuelistas: Pablo Sorozábal (1897–1988), who brought influences from German expressionist theatre into the genre and Federico Moreno Torroba (1891–1982), who followed a more traditional style.¹²⁴ New zarzuelas were constantly appearing (though not at the same rate as in the glory years of género chico) and several became instant classics. For example, Doña Francisquita had sixty performances in Valencia alone during the five months following its premiere.¹²⁵ Within another four months, it had been performed three hundred times throughout Spain.¹²⁶ Other zarzuelas enjoyed similar success, including the previously discussed La del soto del parral, Los gavilanes and El huésped del sevillano, as well as Sorozábal’s Katiuska (1931), La del manojo del rosas (1934) and La tabernera del puerto (1936) and Moreno Torroba’s Luisa Fernanda (1932). And there were some surprises. Even though early predictions for the success of La del manojo de rosas were not optimistic, Sorozábal claimed in his 1986 memoir that after its premiere not a month went by without at least one performance of the work, which led to more than 15,000 performances over the following decades.¹²⁷

These new zarzuelas appeared alongside earlier works such as Gigantes y cabezudos, La revoltosa, La verbena de la Paloma, Marina and La tempestad. Notably, some of the older, more classic zarzuelas such as Asenjo Barbieri’s El barberillo de Lavapiés and Jugar con fuego were rarely performed,¹²⁸ although they were still seen as staples of the genre. (These works would not be revived until the 1950s.) Furthermore, several zarzuelas that were popular during the 1920s and 1930s (e.g., El bateo, La reina mora and El anillo de hierro) remain obscure nowadays. It is likely that recordings played a part in this, for these works were recorded less often than others. Recordings, therefore, did not simply mirror what the repertoire was at any given point in time but also contributed to shaping it.

One consequence of the gradual standardization of a zarzuela repertoire was that singers did not have to learn new works at the same pace as did their earlier counterparts in género chico. Furthermore, they could afford to specialize in a handful of roles. For example, Vendrell claimed that he made his career on the basis of five works,¹²⁹ while Redondo kept an active repertoire of twenty-five works.¹³⁰ These were only fractions of what performers would have been expected to know a generation earlier. Thus, many of the singers we hear in recordings from the early electrical era were singing roles they had performed hundreds of times on stage, whereas this level of familiarity with the part is not always the case for some of the singers heard on pre-electrical recordings.

It was also in the mid 1920s that recording companies systematically started to record and release albums that included the principal musical numbers from a particular zarzuela. These sets consisted of six, eight or ten sides sold together. Whereas it is difficult to pinpoint a direct cause-and-effect relationship between the appearance of such albums and the advent of electrical recordings, it is likely that the increased fidelity and possibilities for aesthetic enjoyment achievable through electrical recordings made such releases more appealing to consumers who were becoming more accustomed to solitary listening. Many zarzuelas that had successful stage premieres were recorded almost immediately thereafter in this album format, some of them several times within a few years. This was the case, for example, with both Doña Francisquita¹³¹ and Luisa Fernanda.¹³² These were not complete recordings by any means. Dialogue was omitted, as were some of the less important numbers and those that would be the most difficult to record, such as ensembles requiring multiple singers.

Although the production of albums that included selections from a single zarzuela was presumably driven by commercial rather than artistic reasons, it eventually had an impact on performance practice. Record companies did not just piece together sides that a few years earlier would have simply been sold separately. Evidence from the recordings themselves suggests that the idea of the album as an artistic artifact (and not just a commercial product) was starting to develop at this time. Some singers were beginning to think beyond individual numbers in the studio and were developing performing and interpretive strategies that would shape the evolution of their characters throughout the album, just as they would likely have done on stage.

One particularly illustrative example of this interpretative approach can be heard in the multiple recordings of Luisa Fernanda that were made in the early 1930s. Act 2 of this immensely popular zarzuela opens with the well-known ‘Mazurka de las sombrillas’, in which the people of Madrid flock to the hermitage of San Antonio de la Florida to celebrate the annual romería. Damiselas (young women) and pollos (young men) flirt with each other nonchalantly. In the midst of the choral revelry, Javier and the duchess Carolina enter and sing a short duet. Throughout the whole scene and subsequent developments, Carolina – a staunch monarchist – attempts to present herself as a non-aristocrat who understands and even participates in several of the events in the life of the pueblo. Some recorded performances of the duet attempt to show this aspect of her persona through performance practice. Ángeles Ottein, in her 1932 Odeón recording, avoids portamento in ways that sound very deliberate,¹³³ especially when compared with her performances of other numbers in the zarzuela, for which she does use portamento.¹³⁴ Furthermore, her contemporaries Laura Nieto,¹³⁵ Sélica Pérez Carpio¹³⁶ and Regina Zaldívar¹³⁷ all employ portamento in the ‘Mazurca’ (see Figure 6). Ottein, therefore, stands out for her deliberate avoidance of portamento as a means to show her solidarity with the pueblo. These other performers, however, employ other expressive techniques to show their connection with the ways of the pueblo. Nieto’s interpretation, for example, features a very distinctive spoken quality, with clear diction and a brisk, excited tempo that avoids the tempo modifications typical of the time. In all four recordings, Javier and the chorus (made up of young men and women from the upper classes who, like Carolina, have come to the festival to experience an aspect of pueblo life) make extensive use of portamento. This prominence makes Ottein as Carolina’s avoidance of the technique all the more obvious and perhaps even comic, as the other outsider characters do not seem as over-keen as she is to fit in.

Figure 6 ‘Mazurka de las sombrillas’, bars 83–91, from Federico Moreno Torroba’s Luisa Fernanda, as written (and performed by Ángeles Ottein) (6a) and as performed by Laura Nieto, Sélica Pérez Carpio and Regina Zaldívar (6b)

Beyond these interpretive uses of portamento, which require comparison between different numbers from the same zarzuela or multiple performances of the same number to become apparent, individual recordings also show some substantial shifts in performance practice when compared to earlier cylinders and discs. Indeed, the 1920s saw what emerges as the greatest documented change thus far in how portamento was used in zarzuela. By now, portamento had become a fundamental vocal technique for singers. It appeared more often but also more discreetly and played a significant role in defining the decidedly more legato approach that characterizes recordings of this era. At the same time, in some contexts it remained a device that could be used intentionally for specific expressive purposes. These roles of portamento become increasingly differentiated, not just in terms of how the portamento was executed but also in terms of how often it appeared. The former type, which I call portamento-legato, became more common, whereas the latter one, which I call expressive portamento, became rarer, subtler and more streamlined.

Unlike the uses of portamento discussed earlier, portamento-legato was applied to smaller intervals (e.g., seconds and thirds) and often appeared in all or most of the smaller intervals in a phrase. In recordings, it is often performed in a non-emphatic manner, which makes it seem more like a default or background consideration rather than a precise expressive gesture used to highlight a particular dramatic moment.¹³⁸ Singing manuals from the electrical era suggest that this awareness of portamento as an aid to legato had increased in the intervening decade. The famed zarzuela tenor Emili Vendrell and the Spanish operatic tenor Hipólito Lázaro (who occasionally performed in zarzuela) expressed similar views in their books on singing technique. Lázaro recommends singing a legato four-note scale, using portamento to connect the notes, in order to help develop the passaggio¹³⁹ and Vendrell describes a similar exercise.¹⁴⁰

During the early electrical era, portamento-legato can sometimes be heard on larger ascending and descending intervals. One outstanding example is the descending fifth on the word ‘imagen’ (image) in Emili Vendrell’s recording of Fernando’s romanza ‘Por el humo se sabe dónde está el fuego’ from Doña Francisquita (see Figure 7). The portamento is understated, almost like a natural consequence of Vendrell emphasizing the highest pitch (on ‘ma’) of the three-syllable word, which he executes in an elegant, controlled manner.¹⁴¹

Figure 7 ‘Por el humo se sabe dónde está el fuego’ from Doña Francisquita, bars 43–44, as performed by Emili Vendrell

Expressive portamento, on the other hand, showed more continuity with previous practices. It was typically applied to larger intervals, descending rather than ascending, and appeared at the ends of phrases rather than at the beginnings. It was used less frequently than in earlier eras and was often executed in ways that resonate better with our modern sensibilities, namely with more consistent breath support and with a quicker and more agile motion. Furthermore, a new general practice emerged: an ascending portamento would be followed by a descending one, particularly in cadential gestures where the middle note receives additional weight through a fermata, whether notated or not. This idea is clearly heard in Mercedes Capsir’s recording of ‘Pensar en él’ from Marina made in 1930¹⁴² (see Figure 8).

Figure 8 ‘Pensar en él’, from Marina, bars 87–88, as performed by Mercedes Capsir for Columbia Graphophone in 1930

There are also instances in which, instead of sliding, the singer clearly articulates every diatonic pitch between the top and the bottom notes. This practice appears in contexts that otherwise would have been ripe for portamento, such as on a descending fourth at the end of a phrase (see Figure 9). It makes sense, therefore, to consider these instances as variations on the expressive portamento, one which perhaps could be regarded as a safer option for singers who did not feel secure in convincingly executing a slide.¹⁴³

Figure 9 Quartet from Agua, azucarillos y aguardiente, bars 38–40, as written (9a) and as performed (9b) by Salud Rodríguez for Gramophone in 1929

As in the past, portamento did not operate in isolation. Changes in its usage can only be understood within the context of broader shifts in the genre’s overall performance practice. It was at the same time that electrical recording was being introduced that the mode of vocal production common in zarzuela recordings nowadays was also emerging. This is characterized by a high soft palate and a low larynx and, consequently, more constant vibrato and less naturalistic diction. In other words, it is at this time that zarzuela recordings begin to sound more and more like opera recordings. This matches our perceptions in 2024 of how zarzuela should sound. Some forms of ‘low soft palate’ technique with a more nasal quality, however, survive among cómico (tenor cómico, tiple cómica) singers.¹⁴⁴ Vibrato was becoming part of the default vocal production, but some recordings reveal, like snapshots, that this was a time of transition. The tenor Delfín Pulido, for example, shows a remarkably consistent vibrato in his 1930 gramophone recording of María la Tempranica, while his partner Felisa Herrero offers a more selective use of vibrato that hearkens back to the time of earlier recordings in which vibrato was an expressive choice applied on selected notes.¹⁴⁵

When it comes to portamento, there are also examples that can be labelled as transitional, for they show how old practices coexisted alongside new ones. In his 1924 recording of Fernando’s romanza ‘Por el humo se sabe dónde está el fuego’ from Doña Francisquita, Juan de Casenave makes his portamento-legatos rather subtle, almost like a spontaneous attempt at mimicking through singing some of the inflections of spoken speech at this particular emotional moment.¹⁴⁶ Vedrell, in his recording of the same romanza from 1927, does something similar.¹⁴⁷

By contrast, Juan García (Parlophon, 1930)¹⁴⁸ and Pepe Romeu (Columbia Regal, 1930)¹⁴⁹ are less subtle in their deliveries. To our modern ears, García’s and Romeu’s performances might sound overemphatic and even excessive, while Casenave’s and Vendrell’s come across as more palatable. The four tenors are all from basically the same generation (Casenave was born in 1888, Vendrell in 1893, García in 1896 and Romeu in 1900), so the differences in performance cannot be attributed to their relative ages. In fact, it is the two older singers in the group (Casenave and Vendrell) who embrace the newer styles. What the recordings suggest, therefore, is that younger singers were at this time experimenting with both older and newer practices. In this particular example, we might ask ourselves whether the fact that the older singers had played the role on stage,¹⁵⁰ while the younger ones had not, may also be significant. Casenave and Vendrell might have been more aware than García and Romeu that the introspective nature of the romanza suggested a quasi-spoken approach, which they emphasized through a more standard portamento-legato.

Vendrell, Casenave, García and Romeu were part of a generation who started their careers around the same time that electrical recording was emerging on both sides of the Atlantic.¹⁵¹ This allowed some singers, such as Vendrell and Redondo, to experiment with new recording technologies from the very beginning of their careers. This generation of singers was therefore different in some respects from earlier ones and this difference likely contributed to the shifts in zarzuela singing at the time. Vendrell’s writings, which include memoirs and singing treatises,¹⁵² show that the singer was keen to present himself and his contemporaries as members of a dynamic, innovative generation intent on changing the way zarzuela was performed, one that left behind the alleged bad taste of past generations. Many of these younger singers had operatic leanings, for some trained as opera singers at conservatoires (e.g., Felisa Herrero and Mary Isaura), while others received scholarships to study in Italy (e.g., Cora Raga and Marcos Redondo) and began their opera careers there before returning to Spain.

An important difference between these singers and those of previous generations is that general musicianship constituted a more prominent role in the younger singers’ training. Many of them read notation fluently, which was not always the case with their predecessors, who usually had to learn on the job.¹⁵³ Vendrell himself insisted that all singers must be able to read music.¹⁵⁴ Composers were well aware of this shift and its implications for performance practice. They were not only writing some vocally demanding music, often influenced by operetta and verismo, but also producing scores that were more detailed in terms of expressivity.

Central to this discussion, slurs were often used to indicate pairs of notes between which portamento was recommended. The absence of portamento in such instances in period recordings, however, suggests that such slurs were seen as encouragements, not mandates, and were ultimately optional. This also implies that fidelity to the score, while increasingly important, still carried with it a significant degree of flexibility.

Composers often ran their own production companies and therefore worked closely with singers. They would have seen and heard first hand how new expressive codes were developing, how they interfaced with existing conventions regarding portamento and how a new generation of singers was balancing the two approaches. They would also have had ample opportunity to reflect on how to translate these practices into notation, but always with the awareness that singers could (and often would) still resort to their previous approaches to interpretation when it came to departing from what appeared on the page. Zarzuela singers of the 1920s and 1930s, it becomes clear, were highly skilled professionals. They were musically literate and keenly aware of unwritten conventions. Their use of recording technologies to benefit their careers and to document their unique performance practices, including portamento, was unprecedented in the genre.

Zarzuela and Zarzuela Recordings during and after the Spanish Civil War, 1936–1953

The Spanish Civil War (1936–1939) disrupted musical life in Spain in profound ways. Many leaders of the musical avant-garde active in the 1920s and 1930s, such as Salvador Bacarisse and Rodolfo Halffter, were forced into exile. This precipitated a leadership crisis in the first few months of the war, one that forced the veteran composer Joaquín Turina into the position of Comisario de Música. Good orchestral conductors were seen as few and far between, a dilemma that constituted a serious obstacle to the continuation of musical life.¹⁵⁵ In the realm of zarzuela, however, the disruption was felt to a lesser extent.

During the war, zarzuela continued to be performed extensively by both Republicans and Nationalists, which is a testament to how zarzuela could resonate among a range of national identity concerns, regardless of specific ideology. Performances typically relied on the established repertoire and not on new works.¹⁵⁶ On both sides of the political divide, companies needed to be able to quickly mount productions that would both be financially solvent and improve public morale.

Proof of how popular zarzuela was on the Republican side is the fact that some of its cities nationalized their theatres to keep them going. Singers and in fact all theatrical workers were paid a modest salary from the government. Under this system, a primera tiple would often earn the same as a driver. In Madrid, which remained in the hands of the Republicans until the final days of the war, numerous beneficio performances were organized with the aim of collecting donations for various pro-Republican causes. Musical life in Madrid was fast-paced and often chaotic. In order to boost attendance, performances were often advertised as featuring certain very popular singers, who would then fail to appear. Sometimes, the same singer was advertised as performing in two different theatres on the same day at the same time.¹⁵⁷ While such announcements prove the unreliability of individual press reports, they collectively show that zarzuela remained extremely popular with audiences in Republican Madrid. Zarzuelas flourished alongside lighter genres with similar popular appeal, such as revistas. In fact, zarzuelas and revistas sometimes appeared on the same programme.¹⁵⁸

On the Nationalist side, the singers Emili Vendrell, Hipólito Lázaro and Marcos Redondo, all of whom were well established in the zarzuela scene, found themselves in Barcelona when the theatres there fell under the control of the anarchist trade union CNT (Confederación Nacional de Trabajadores). They were given a salary of fifteen pesetas per performance (in peacetime, they could earn as much as one thousand¹⁵⁹) and were having to sing physically and vocally demanding runs in Barcelona, in various towns throughout Catalonia and on occasional international tours.¹⁶⁰ Such activities, organized by the Catalan government, were designed to keep morale high on the home front. So that companies could cope with the massive demand for performances, repertoires were pared down to a small number of extremely popular works, both old and more recent. These included Marina, Doña Francisquita, La rosa del azafrán, La tabernera del puerto, La del manojo de rosas, Bohemios, El dúo de la Africana and El cabo primero.¹⁶¹ At the same time, the Catalan government was showing interest in revitalizing to a certain extent the more prestigious Spanish operatic repertoire through performances of José María Usandizaga’s Las golondrinas,¹⁶² Tomás Bretón’s La Dolores, Enrique Granados’s María del Carmen and Manuel de Falla’s La vida breve.¹⁶³ Furthermore, the Nationalists reconceptualized the social role of radio and for the first time in Spanish history radio was being used as a propaganda tool. This led to zarzuela often being featured on the leisure broadcasts of Radio Nacional, alongside opera, Spanish art music and standard classical repertoire.

Central to the present discussion, very few, if any, new recordings of zarzuela were made by either side. While there certainly would have been constraints to producing such recordings, particularly in terms of international circulation, it is likely that both Nationalists and Republicans were sceptical as to recordings’ potential value as propaganda. Live performances would have no doubt provided a sense of community among those in the audience who were coming together in the same physical place and listening to live radio broadcasts would have engendered a sense of connection among those listening to the same programme at the same time all over Spain, an ‘imagined community’, to use Benedict Anderson’s term.¹⁶⁴ Recordings, however, because they lacked the live quality that could bring people together, might have been regarded as less useful for identity-building purposes in a time of war. The act of solitary listening that had been developing in previous decades had lost its foothold.

After the Civil War ended in 1939, the zarzuela scene re-established itself relatively quickly. Some companies toured the provinces while others settled in the main theatres in Madrid and Barcelona. Signs were emerging, however, that the genre was losing some of its energy as a living art form. Companies were becoming increasingly reliant on past repertoire and the lack of sufficient financial funding and artistic vision were being felt. Productions were all too often regarded as lacklustre and poorly rehearsed.¹⁶⁵ Some of this can be attributed to the uncertainty in which the music profession had to operate in post-war times (as was also the case for other art forms). Younger audiences were choosing other types of entertainment such as the cinema and new generations of composers were no longer interested in creating zarzuelas, particularly since other, more lucrative possibilities were emerging in other areas of popular music.

Although live performances of zarzuela resumed after the war ended, recordings did not. This likely had to do with the lack of vigour of the zarzuela scene on stage and the absence of significant technological innovations that could have sparked either a renewed interest in recorded music or a shift in performance practices. While the continuous production of recordings from 1900 to 1936 (sometimes with the same singer recording the same number several times within a few years¹⁶⁶) often offers crucial evidence to show how certain performance practices gradually changed, for the period from 1936 to 1953, we are, to a great extent, in the dark.

Some important recordings were made in the post-Civil War years, to be sure. These include composer Pablo Sorozábal conducting selections from his successful, yet controversial zarzuela La tabernera del puerto¹⁶⁷ in 1943. Evidence from such recordings suggests continuity rather than drastic change in the use of portamento. For example, in ‘terceto cómico’, the comic characters, performed by the singers Enriqueta Serrano (as Antigua) and Manuel Alares (as Chinchorro) employ portamento of the comic variety described above, while the more serious character Marola, performed by Pepita Embil, does not.¹⁶⁸

Zarzuela on Long-Play (LP) Records, 1953–1958

The next significant milestone in the history of zarzuela recording did not arrive until the 1950s, following the introduction of the long-play 33 rpm record by Columbia in 1948. This technology made it possible to record complete zarzuelas for the first time. From 1953 onwards, several conductors and labels undertook full recordings of zarzuelas. These were made mostly through a series of small Spanish labels (e.g., Hispavox, Alhambra) that were affiliated with multinationals. The recordings were intended primarily for the local market in Spain, although some full-length recordings were also made in the Americas for Spanish-speaking audiences in the Western hemisphere.¹⁶⁹ Since new works were rare, these recordings focused mainly on established repertoire. Some especially popular zarzuelas were recorded multiple times within a short period. For example, no less than three recordings of Gigantes y cabezudos appeared in 1954 in what seems to have been a race to conquer a growing domestic market.¹⁷⁰ Other zarzuelas, such as El rey que rabió, were being recorded for the first time since the early 1930s.¹⁷¹

Significantly, this revival of zarzuela recordings coincided with what can be termed the death of zarzuela on stage. It was no longer a thriving theatrical genre characterized by the creation and production of new works. Neither of the two most popular living zarzuela composers at the time, Pablo Sorozábal and Federico Moreno Torroba, were producing new works to the same degree as they had in the past and neither was enjoying the same level of success as had been the case in the 1930s.

Zarzuela’s new home turned out to be the LP record. Several conductors were central to this revival of classic zarzuela in the recording studio. Especially important were Ataúlfo Argenta, Daniel Montorio, Enrique Navarro, Rafael Ferrer and Ricardo Estevarena as well as the previously mentioned composer-conductors Sorozábal and Moreno Torroba. In the cases of Sorozábal and Moreno Torroba, they mainly recorded LPs of their own zarzuelas, just as they had done in the 1930s with individual zarzuela numbers.

It was Argenta, more than anyone else, who restored the dignity of zarzuela by taking it into homes throughout Spain and indeed around the world at a time during which the genre’s popularity on stage was increasingly in decline. A key figure in the histories of zarzuela and post-Civil War Spanish musical life, Argenta made zarzuela a respectable genre on par with the classical canon. When critically examining his invaluable contribution to zarzuela, however, we must bear in mind that his contemporaries mythologized him and his achievements to a considerable degree. His untimely death in 1958 at the age of forty-four no doubt contributed significantly to his legendary status.

An anecdote from the journalist and critic Antonio Fernández-Cid is particularly illustrative of how Argenta’s reputation as a champion of zarzuela was built. Fernández-Cid claimed that Argenta often told his musicians to perform the accompaniment to the dialogue between Don Sebastián and Don Hilarión in La verbena de la Paloma with the same care and attention as when they were playing Mozart.¹⁷² Zarzuela, in his estimation, was equal to opera buffa as a musical-dramatic genre. Argenta’s supporters also routinely commented on his loyalty to the notated score as fundamental to his efforts as a renovator of zarzuela. He was said to have brought the genre in line with practices of strict adherence to the score typical of the classical canon.¹⁷³

Key to Argenta’s reputation was the fact that he, unlike many other conductors recording zarzuela during these years,¹⁷⁴ had achieved considerable success both in Spain and abroad conducting the standard orchestral repertoire. Several biographers claim this as proof that Argenta was truly committed to zarzuela and was even happy to let his international orchestral conducting career take second seat to his work in zarzuela. He intentionally chose zarzuela rather than being pushed into it because of a lack of other options.¹⁷⁵

Argenta’s status as a champion of zarzuela was also due to his choice of repertoire. In six years, from 1953 to 1958, Argenta recorded around sixty zarzuelas, an astonishing number.¹⁷⁶ The sheer quantity of his output as well as his repertoire choices reflect his desire to establish a zarzuela canon at a time when the genre was not teeming with new works. Argenta’s canon was decidedly composer-centric, with eleven composers taking up almost ninety percent of the list. This explains why his discography includes, besides the hugely popular pieces that had been in the repertoire since the late nineteenth century, several relatively obscure works by otherwise well-known composers, such as Chueca (El chaleco blanco) and Chapí (La patria chica).

In Argenta’s hands, the history of zarzuela thus became a list of composers and works, similar to the way that Western classical music was being canonized at the time. The early history of modern zarzuela – up to the birth of género chico in 1880 – is represented in Argenta’s recordings by three zarzuelas grandes from the composer Asenjo Barbieri, two of which, Jugar con fuego and Los diamantes de la corona, had rarely been heard either live or on recordings during the previous decades. The third work, El barberillo de Lavapiés, on the other hand, was an established classic, even though it was not performed all that often. These three recordings suggest that Argenta was striving, at least somewhat, to reclaim a golden age of zarzuela that modern audiences did not know particularly well. His efforts in this regard were limited, for he did not record any works by other zarzuela grande pioneers of the time, such as Joaquín Gaztambide, Pedro Marqués, Cristóbal Oudrid or Tomás Bretón (in his pre-La verbena de la Paloma endeavours).

Argenta also made numerous recordings of género chico. In doing so, he clearly embraced the views that some in intellectual circles had been expressing since the 1910s:¹⁷⁷ género chico, even though light-hearted and incidental in some ways, featured some excellent music and performed an embodiment of Spanish identity. As was to be expected, Argenta recorded universally successful works that had remained in the repertoire practically since their premieres (e.g., La revoltosa, La verbena de la Paloma and La Gran Vía) as well as others that had achieved considerable success in their time but then fell in popularity (e.g., El barberillo de Lavapiés and El juramento). Argenta therefore provided listening audiences with a broader and more nuanced picture of género chico than what could have been seen on stage.

While it would be anachronistically misleading to speak of a commitment to historical performance practice in Argenta’s recordings of the earliest zarzuelas, an examination of portamento in his recordings can provide further insights into the conductor’s ambition to reshape zarzuela’s past in terms of its present. Zarzuelas such as La viejecita and Doña Francisquita had been regularly recorded since their premieres, which allows us to hear a continuum in terms of their performance practices. Others, like El rey que rabió and La verbena de la Paloma, whose popularity waxed and waned, were also recorded numerous times over the years. These recordings provide us with aural snapshots through which we can hear and understand how the approach to some numbers or roles changed over time.¹⁷⁸

This is not the case with El barberillo de Lavapiés, however. After the 1902 recording of the romanza ‘Canción de Paloma’ discussed earlier, the zarzuela was not recorded to any significant degree for about fifty years.¹⁷⁹ It only returned to the studio in the 1950s with Argenta and his contemporaries. Argenta’s recording was released in 1955¹⁸⁰ and another, conducted by Enrique Navarro, appeared the following year.¹⁸¹ The uses of portamento in these recordings yields several telling insights (see Figures 10–13).

Figure 10 ‘Dúo-tirana’ from El barberillo de Lavapiés, bars 49–51, as performed in the 1956 Hispavox recording conducted by Enrique Navarro

Figure 11 ‘Dúo-tirana’ from El barberillo de Lavapiés, bars 111–12, as performed in the 1956 Hispavox recording conducted by Enrique Navarro

Figure 12 Tercet from El barberillo de Lavapiés, bars 16–17, as performed in the 1955 Alhambra recording conducted by Ataúlfo Argenta

Figure 13 ‘Dúo-tirana’ from El barberillo de Lavapiés, bars 81–82, as performed in the 1955 Alhambra recording conducted by Ataúlfo Argenta

These examples are really not that different in their use of portamento conventions from other recordings discussed in this Element, particularly those following the arrival of electrical recordings. Expressive portamento was applied mainly on larger descending intervals at the ends of phrases, though less often than on the earlier cylinder recordings. We can therefore speak of a performance tradition that was part reconstruction and part invention, with Argenta and his singers not aiming so much to replicate how El barberillo would have been performed in the 1880s – such an approach would be have been largely unthought of at the time – but rather to perform the zarzuela in a way that fit within the genre’s expressive codes as they were understood in the 1950s. On the other hand, the extensive use of portamento in Argenta’s recordings, as well as other performance decisions such as frequent tempo modifications, calls into question his reputation for always sticking to the score. Such liberties are often heard in other zarzuela (and opera) recordings from the decade and are not unique to Argenta’s approach to zarzuela.

None of the singers who sang on these LP recordings completely broke away from past performance practices while they were making their own marks on what might be considered to constitute a zarzuela aesthetic at the time. Some of these singers came from the first generation to take full advantage of electrical recording, such as Marcos Redondo (who in 1954 recorded a new Luisa Fernanda for Hispavox conducted by its composer, Moreno Torroba)¹⁸² and Sélica Pérez Carpio (who played the supporting role of Gorgonia in Argenta’s recording of La revoltosa).¹⁸³ In recordings of romanzas from Doña Francisquita conducted by Rafael Ferrer in 1953¹⁸⁴ and Argenta in 1954,¹⁸⁵ the use of portamento maps remarkably well to what has been discussed in this Element so far. Since Doña Francisquita never left the repertoire, it is likely that Emili Vendrell, who kept singing Fernando until his retirement in 1953, provided such a strong link with past portamento traditions that they almost became conventions. The role of such singers in ensuring a continuity in performance practice was significant, for through their LP recordings they provide aural glimpses of what little we know about zarzuela performances practices in the years following the Spanish Civil War.

Most of the singers we hear in 1950s LPs of zarzuelas, however, were newer to the scene. Argenta preferred to work with the same small group of singers for multiple recordings: Ana María Iriarte, Toñy Rosado, Inés Rivadeneira, Manuel Ausensi and Carlos Munguía. These were all singers with whom he had recently worked on stage.¹⁸⁶ All of them were at the beginning of their careers and many, like Iriarte, were recent conservatory graduates, with minimal or no professional stage experience. This careful choice of singers became a key feature of Argento’s reputation as a renovator of zarzuela. His various biographers collectively agree that Argenta did not want to hire singers who had acquired bad habits and settled for substandard singing technique from years of inadequate, rushed rehearsals and working with poor conductors.¹⁸⁷ These biographies do not state exactly what these bad habits and techniques were, however. Not every older singer who recorded on LPs fit this description, as we can hear in recordings by Vendrell, Redondo, Pérez Carpio, Raga and others.

Several of the younger singers heard on zarzuela LPs from the 1950s also had careers in opera, such as Pablo Civil and Luis Sagi-Vela, the latter the son of noted zarzuela performers Emilio Sagi Barba and Luisa Vela. Some would soon establish international careers, as was the case with Alfredo Kraus, Teresa Berganza and Pilar Lorengar.

These younger singers had more systematic classical training than those from previous generations, which to a great extent translated into the continuation of certain performance practices, including the use of portamento. The change in aesthetics when compared to earlier decades of recorded technologies is perhaps most obvious, though, in the increasingly consistent use of vibrato. Still, some remnants of past performance practices lived on, as is revealed in the 1956 Hispavox recording of Luisa Fernanda conducted by Enrique Estela.¹⁸⁸ Playing the role of Duquesa Carolina, Natalia Lombay’s vibrato is very wide in the upper range, but her middle and low registers reflect the low soft palate placement and chest dominant delivery reminiscent of earlier eras.

When compared with previous decades, portamento technique in the 1950s became more uniform across zarzuela grande and género chico. There are also hints that determining when portamento would be used had also become more standardized. For example, the portamento-legatos on intervals of a second, as discussed earlier, continued to be standard practice. Expressive portamento appears mostly on descending fourths and fifths (and to a much lesser extent, on octaves and thirds), normally at the end of a phrase (though sometimes in the middle) and usually on a word the performer may have wanted to emphasize. Improvements in recording technologies as well as in the vocal training of singers allows us to hear greater nuances than had been the case previously. This confirms that the stylistic trends heard in pre-Spanish Civil War recordings continued into the 1950s. While each singer’s approach to executing portamento certainly remained an individual choice, portamento in general was now being performed quickly rather than slowly and in a more understated way than was the case in the very early recordings. Overall, this practice follows contemporary developments in other genres while resonating more readily with present-day tastes and sensibilities.

Conclusion

The plethora of studies that make use of early recordings as sources for performance practice have given us much more knowledge now than we had ten or twenty years ago about how performing styles (for specific instruments, genres or geographies) changed over the first decades of the twentieth century. Tackling the question of why performing styles change, however, is more challenging and is not always explicitly addressed in such studies. Beyond attributing shifts to nebulous matters of taste, some authors attempt to offer answers. Mark Katz, as discussed earlier, coined ‘phonograph effect’ as a catch-all term to explain how recording technologies influenced the way musicians performed in live situations¹⁸⁹ and Daniel Leech-Wilkinson hypothesizes that changes in performance practices towards the middle of the twentieth century might have been motivated by societies becoming more self-conscious and cynical after the Second World War.¹⁹⁰ Interestingly, Katz and Leech-Wilkinson’s answers fall on almost opposite ends of the spectrum: from the eminently practical, logistical and mechanical to the overwhelmingly abstract.

In this Element, I have attempted to illustrate how early recordings can be used not only to locate and describe change but also to explain how and why it came about. My task has been considerably facilitated by zarzuela being primarily, though not exclusively, composed and recorded in one specific place: Spain. The study of zarzuela, though, truly calls for transnational approaches. It was shaped throughout its history through dialogues with musics from elsewhere and it circulated in Spanish-speaking countries outside Spain. Zarzuela’s production and consumption through the medium of recordings, however, was more localized than was the case for, say, opera and orchestral music (with some notable exceptions). This geographical limitation has made it easier to identify and pin down the various contexts and spheres likely to have influenced the performance practices associated with portamento.

These practices start with the recording technologies themselves in terms of both their technological development over time and the commercial practices they engendered. Such shifts in turn resulted in new ways of consuming and listening to music and in new discourses on music and sound. Also important were changes in the zarzuela scene itself. With some of them, such as the working structure of the music profession and the training of singers, it might be relatively simple to establish cause-and-effect relationships; such matters, however, are often closely connected to broader issues concerning the social status of zarzuela at any given point in time and its perceived connections to national identity.

The challenge is to determine the extent to which such factors might be heard in recordings. There is a real danger to quasi-mechanistically relate every performance element to a contextual development. Such actions risk losing sight of key aspects of these recordings and the performances they capture. Beyond performance conventions and technological constraints, there was still ample room for personal preference and for the randomness that is inherent in the interactions between musicians and recording technologies. Taken together, these factors tend to delimit a range of expressive possibilities for portamento rather than establishing a set of hard and fast rules. What we hear in these recordings are some of these expressive possibilities. Especially for the earlier years, it is easy to see how the picture remains incomplete, but a comparison of recordings, as well as consideration of their cultural and social circumstances, can certainly help us make sense of practices that on first hearing might sound hybridized, changeable and multifarious.

One of the key factors in this Element has been the notion that zarzuela’s requirements and priorities were dominated to a considerable extent by the need to convey text expressively. The centrality of language powerfully connects portamento performance practices in zarzuela with those in other genres. Portamento has often been described as an attempt to translate aspects of speech into singing and the present study offers a more detailed look into how this happened in the realm of zarzuela.