Barry Kay, internationally acclaimed stage and costume designer, photographer, born 1932 Melbourne Australia, died 1985 London England

On designing for ballet today


Barry Kay talking to Alexander Schouvaloff




I remember that my first real interest in the theatre was due to the Ballet Rambert Company which came to Australia when I was fifteen. I was enormously impressed. It was at a time when they had reached an especially high point artistically. This is what gave me my first thoughts about working in the theatre.

Ballet in particular?

Yes. At least at that stage. Until then my main preoccupation was music – piano and composition. I was interested in composition from very early on. From the age of fifteen I began to develop in a new direction – towards painting. So with a little persuasion my parents sent me to art college in Melbourne. It was not very long after that I managed to go to Paris to study painting at the Académie Julian. But for me the most exiting part soon became the theatre. This was Paris shortly after the war and it had burst forth with some of the most remarkable stage designing to have been seen since the Diaghilev era. It was from that moment that I really decided this was what I should do. The Ballets des Champs-Élysées was the focal point with its productions by Roland Petit. I think it caught a lot of our imaginations. I believe that it did in the case of Kenneth MacMillan who felt a similar reaction at the time. I managed to see everything they did many times over. Also I met and came to know designers like Clavé and others, which in some ways helped more than the academic work I was being taught at school.

I then returned to Australia for a short time. and began working with Walter Gore, who had formed a company out there with his wife Paula Hinton. I designed several ballets for them which really gave me a basis for the work I was to commence in Europe.

"On designing for ballet today" is the tile of a talk between Barry Kay and Alexander Schouvaloff. It took place in 1980 and constitutes one chapter of the richly illustrated book "Designing for the dancer" [1].

Alexander Schouvaloff was Curator of Theatre Prints at the Victoria & Albert Museum (1974-1989), as well as Founder and Curator of the London Theatre Museum, which was reintegrated into the V&A Theatre and Performance Collections in 2007.

For ease of reading, questions posed by Alexander Schouvaloff are set in italic, responses by Barry Kay in roman.

Why didn't you want to stay in Australia?

Because I felt there was very little further advancement I could make in the Australian theatre. What I had seen in France had left a great impression and the Australian theatre seemed both dull and limited; not what it has become today.

Besides, there was an unfortunate snobbery about local talent which denied progress to those who had not travelled or achieved success abroad. It's not that I particularly responded to that pressure but I needed to develop in a more demanding climate. It seemed that the only way that I could hope to expand, was to go abroad.

This is how I came to London where I worked for a short time with the Western Theatre Ballet [now Scottish Ballet; ed], a very interesting and adventurous small company. This followed with plays which I designed for the Old Vic and the Royal Shakespeare Company. Around this time I meet Kenneth MacMillan with whom I have since had a close partnership in designing many of his ballets. It has been a considerable influence on my work.

What do you mean by 'a close partnership' between designer and choreographer?

Well it doesn't always happen. I think it is something which develops through a long period of working together. The approach has to be complementary so that each one knows the direction in which the other is going without necessarily having to say it. I have always experienced this with MacMillan and also with Nureyev. Such a working partnership requires considerable perception on both sides so that not everything has to be verbalised. I remember with MacMillan, right from the beginning, we often had very little to say and the understanding of the work grew as we went along. But as we have begun more complex work, of course a great deal of investigation and discussion is needed in order to achieve a solution to the problems which arise on such productions.


IMAGES OF LOVE, ballet, costume designs, drawing by Barry Kay, choreography Kenneth MacMillan, The Royal Ballet, 1964

Images of Love

Costume drawings

The Royal Ballet
Royal Opera House
World première 2 April 1964


Does the choreographer always take the initiative?

No. It's possible that various ideas may come from the designer. This happens quite often. MacMillan in particular has a gift of being able to take a given idea, to absorb it, and to integrate it so well into a production that you can't tell from where it has originated.

On the other hand he frequently offers the most excellent design suggestions. By the end of a production there is such a fusion of ideas that you could never tell where they all began.

You have decided, the two of you, to do a ballet and the ideas start springing out. In the practical sense what happens?

Right from the beginning?

Right from the beginning.

Can I say one thing? You have been asked by the choreographer because he knows you are the right choice for a particular production. This is one of the reasons, as I have said before, that there isn't all that much need for verbalising from the beginning. In any case you are likely to be in agreement because of the similar viewpoints which have brought you together in the first place. We usually begin with a general discussion about the production. Then I put very rough ideas together which have probably formulated during our initial talk. Invariably I need to go through a long digestion period. Not necessarily putting ideas down on paper, rather thinking and mulling them over. But I have to do this for quite a long time.

How long? Months or weeks?

It could be weeks but usually longer. In the case of Isadora which we are working on now, it was many months as it is such a complex production. It did require a lot more thinking in order to arrive at a total – so that I didn't see it in parts. This is always the danger in a large work, a full-length production with numerous scenes – that you will see it in sections. You mustn't. No matter how sketchy, it is essential to visualize a total from early on.

Next come quite rough notes. I never make anything too elaborate at the beginning in order not to frighten myself – so just rough sketches that can be torn up – and when I feel that there is an idea emerging, I go back to a discussion with the choreographer. More often than not I find that we are much agreed, that we have been thinking along similar lines.

This was much the case when I designed Kenneth MacMillan's production of The Sleeping Beauty, for the Deutsche Oper in Berlin. He was then Director of the Ballet Company and had invited me to Berlin in order to exchange ideas on the way in which the production might be mounted. I had thought of transposing the story from France to Russia, making it closer to the music of Tchaikovsky. Traditionally it has always been set in the French Courts of the 17th Century with the awakening taking place in the 18th Century. My proposal was to set it all a century later commencing with the Court of Catherine the Great. In this way the court's awakening would be contemporary to Tchaikovsky's score.

To my surprise, when I arrived in Berlin, I found that we had been thinking along very similar lines, especially over repatriating the work to Russia.

arrow up


THE SLEEPING BEAUTY, Prologue Set Design, drawing by Barry Kay, choreography Kenneth MacMillan, Ballet of the Deutsche Oper Berlin, 1967

The Sleeping Beauty


Set design, drawing
Carabosse centre stage

Ballet of the
Deutsche Oper Berlin
Première 8 October 1967


We'll talk about costume in detail in a moment, but nowadays designers tend to do all the production, both the scenery and the costumes, don't they?

Yes, on the whole.

But would you be against one person designing the scenery and another the costume?

Well it depends. If the work is stylised with inventive sets and costumes, then certainly there needs to be a total cohesion between the two elements – this is something requiring a single vision. However, in a production which is more realistic in style, I can imagine that it wouldn't be necessary. For instance, if you had historical costumes which fitted into an equally realistic setting, then I could imagine it possibly being designed by two people. The only thing is that I have never done that myself and I don't think I would ever want to do it that way. To give another example, the production of Isadora which I am currently designing, requires costumes of a documentary nature and a contrasting set. To bring the two together successfully, I need a grasp of the whole concept as there is a fine balance to be considered when you are combining such disperate elements.

So how do you begin?

I usually commence with the sets. I think I've hardly ever begun with the costumes. The sets act as a base from which the rest develops. Usually I find it easier to work with rough bits and pieces – scraps of card or wood, just to see something in the round. This is in order to have some idea of how I might use the space. I have always felt more comfortable working in a sculptural way rather than going direct to the drawing board. I tend to alternate between building things very roughly (to see a shape within the given space) and making drawings which further those ideas. I develop them simultaneously. But in formulating a designs, the most important element remains the sculptural or architectural one, as you've probably noticed from my rather constructivist sets.

As I am dealing with something three-dimensional, I find it confining just working it out on paper. It would all remain too flat when starting to build the actual model.

Of course I'm talking about the proscenium stage. But this applies even more with the theatre-in-the-round where you're obliged to conceive the design in a three-dimensional way. I remember when I began designing for the international opera houses that they were still used to using flat painted scenery for the ballet. Ballet design in the classical tradition continued to be influenced by Diaghilev's use of easel painters. Other developments which had taken place during the twenties were forgotten. When I produced my first design for Covent Garden – Images of Love, it was not easy for them to accept a completely built set. And to begin with, I encountered even more resistance to change from opera houses on the Continent.

It's this structural side to design that has largely motivated my continuing interest in ballet. The limitations of ballet design are unique and I suppose that I've always found these limitations a challenge. You are faced with this dance space that must be kept quite free while at the same time attempting to create a convincing location. The more realistic the environment, the greater the problem. There are various devices which I have worked on and developed over the years, particularly in the use of the space overhead.

These constructions above the dance area serve to describe what might exist upon it without actually having to be put there, except perhaps for a few props. Quite often you find variations on ceilings and other more abstracted structures in my sets.

arrow up


ANASTASIA, ballet, 3-act verion, Stage Model Act I, Design Barry Kay, Choreography Kenneth MacMillan, Royal Ballet, Royal Opera House, London 1971


Three-act version, Act I

Set model, scale 1:25

The Royal Ballet
Royal Opera House
World Première 22 July 1971

Victoria & Albert Museum
Theatre and Performance Collections


And it's into this setting that you put the dancers?

Yes. Coming back to costume! We've veered a great deal, but it does relate to costume, for besides any other consideration I also do this so as to bring the costumes into as close a relation with the sets as possible, whether the purpose be for unity or contrast.

What are the particular problems of dance costume?

The biggest one lies in the fact that there are enormous limitations as to what you can do on a dancer's body. That's one one of the first problems you face. In the classical ballet there is one of the greatest restrictions. The tutu. What do you do with a tutu? It's a shape which you can only vary slightly – so basic that there is just the question of what pattern will go on top of it. Early on, I actually learned a great deal from Haute Couture. I used to go frequently to fittings at various couturiers in order to study how they constructed costumes. Their approach to it was quite architectural, which had greater appeal to me, than what I had seen of ordinary dress-making methods. I felt I could learn more from the couturiers about the discipline of proportion, line and form, and I've tried very much to bring that into the theatre. They have a special finesse in construction. It's not the detailing that goes into the making so much as the attitude towards the building of a costume. It's something which influenced me greatly. Equally important for me is drawing. Not that I think it is necessarily a prerequisite for good design, but I think it is a great help in understanding proportions and being able to deal with the body to best advantage. Not only that, a good drawing is a greater incentive to the maker in many respects. It is something I chose to study and to practice from quite early on. In any case, I have always been conscious of the fact that most of the best designs in the past have been produced by those who were also excellent draughtsmen – from the Bibienas to Bakst.

How detailed does the drawing need to be?

Well I have a sort of shorthand in my drawings so that the design can be understood immediately by the maker. I don't need to make extra working drawings. But that's taken years of working out. A direct approach with a limited amount of detailed explanations.

So then the drawing is agreed upon, the final design is agreed by the choreographer?


And then it goes to the workroom? Do they make patterns first, how do they proceed?

Yes. They usually make up toiles, which is also similar to couture. From those toiles they begin work on the actual costume. One of the major factors of course is the understanding of what material will do in movement. In that respect it is vital that the designer has a thorough knowledge of material. In order to make a successful ballet costume, it probably comes before anything else, because if you've chosen the wrong material then, no matter how good your cut is, it's not going to work at all as a dance costume.

You therefore detail the material the costume is to be made in, and all the trimmings?

Yes, all the materials have to be selected and this is a very long process. When all the designs are assembled you begin to select, largely through pattern books or materials which happen to be available in the theatre's stockroom – they usually keep quite a large stock of materials, but you always have to obtain the bulk of it from other sources. While I'm designing I have a very clear idea all the time about the type of fabric I intend using. It's by force of habit – not something I am all that conscious of; it's just there while I'm working on the designs.

Do you find those materials on he market?

Sometimes they have to be imported.

Do some of the designs require special fabrics?

Very often. I don't always use material straight. That is to say, sometimes it needs to be combined with other materials, overlays and appliqués to destroy that 'dressmakerish' look that material can often have on the stage unless it is treated in various ways. Sometimes it's with dyeing. The texture is very important on stage, particularly in relation to how it reads at a distance.

What materials are generally used for dance?

That varies considerably. Of course net is always used for the classical tutu. But for other costumes, it all depends on the quality of movement required. There are certain materials such as chiffon, silk and organza which move better than others, but on the whole, in terms of movement there is a rather limited choice. Numerous variations exist within the same range of fabric, but in actual fact the number of differnet materials available is not great.

Speaking of limitations, there is another which is considerable – the narrow possibilities that exist in constructing a dancer's costume. Not only is it a question of adapting to movement or being familiar with the way in which the body works. To realise any successful costume, a well-structured costume, there must be an understanding of the divisions which most enhance the dancer's body – which lines and their placement look best both in cut and ornament. A designer can bring his individual style to solving this problem as well. One of the most common mistakes can arise when a painter or sculptor, not previously experienced in the technicalities of costume making, is commissioned to design a production. There is often a tendency to pin a drawing directly to the body – to project the two-dimensional picture straight onto the figure without fully taking into account that there is also a back and sides. This is one of the reasons Diaghilev employed Chanel to aid as an adviser on various occasions, as painters were frequently invited to design for his company.

Yet another aspect lies in the relation of the design to the individual who is wearing it. I find that there is often a need to allow for a margin of change, but this is usually in the case of soloists. It's really a matter of knowing how to retain the essentials of a design without imposing something which doesn't suit the dancer. Sometimes this involves problems in movement.

arrow up


Five preparatory sketches, preliminary drawings, costume designs by Barry Kay, for character dancers in Anastasia 3-Act version, Act II, ladies fashion between 1910 and 1920, The Royal Ballet, 1971, Royal Opera House, Covent Garden, Choreography Kenneth MacMillan


You get to know this movement?

Yes, or by knowing the choreographer's work. It's not always necessary to see the choreography at first, although more often than not I would say it is essential. It depends. If you take one of the classical productions such as Sleeping Beauty or Swan Lake, you are dealing with a known quantity, but with an original work there is likely to be a wider range of alternatives when it comes to style of presentation. However, this depends on the choreographer's approach to a new work and whether he has formulated a precise visual concept which he needs interpreting or whether he requires the additional vision of the designer with whom he has chosen to work.

Have you ever found that you've made a complete mistake?

Once, only once. But it was certainly complete as it involved the entire set of costumes for the ballet Cain and Abel which I designed for the Deutsche Oper in Berlin. Unfortunately I hadn't seen the choreography because at the time I was also working in London. I had completed the designs and been over quite briefly for a number of fittings but without having seen rehearsals. I arrived for the dress rehearsal, we sat there and the curtain went up. There were all these costumes glaring at us, entirely wrong for the choreography. Besides the surprise, I remember it was quite mesmerising. Anyway we scrapped them and overnight came up with something completely different. The wardrobe was remarkably helpful. It was somewhat traumatic, but we re-did it. I must say that's the only time it has ever happened. [2]

Was that because you had not been in touch together?

It was a mistake because I should have been on the spot during most of the preparation. This is why I usually devote the maximum amount of time to one production without other distractions. On this rare occasion I was obliged to divide my time. Needless to say I don't any more.

Do you find at all that some costumes have to be adapted once they are made, that they have to be changed during performance? The dancer may be happy in dress rehearsals but...

Not really, because I find that I solve most of the problems which arise in the fitting room with the dancer. An experienced dancer is usually quite confident in what you're doing, or have designed. I rarely have difficulties in this direction because I'm always conscious of the dancer's problems and incorporate this into the supervision of the work, so that – unless I find there's some psychological barrier – it's usually straightforward.

One of the most excellent people to design for is Margot Fonteyn as she possesses perfect proportions. I have never met another dancer, either male or female, who has such proportions, so that when you're at a fitting it's patently clear where things have gone wrong. Her body seems to make everything look clumsy if there's the slightest mistake but immaculate when it's right.

Do you approach every ballet in a similar way?

No, not at all. Each time it's a new solution.

When you're designing a classic are you influenced by previous productions? Are you aware of previous solutions?

Yes I am. But I'm not influenced. I mean the influence could be to go in quite another direction. With Sleeping Beauty for instance, it was an interesting example. Having had a musical background, I have always given special attention to the music. So I took the cue from Tchaikovsky's score and thought there was a curious discrepancy between the French Court in which the production is usually set and the 'Russian-ness' of the music. I listened to the music many times over and felt that somehow there must be a way of getting closer to this 'Russian-ness'. And then I began to look at a lot of material, architectural mostly, and saw a way of bringing it together in a Russian Court setting. Now there was one thing I knew which MacMillan wanted to do from the beginning. He wished to create a big contrast between the people of the Court and the fairytale figures in oder to have two distinct levels in the ballet and not one as is usual. In this way it would lend an unexpected weight to the production. I've always found a certain barbaric quality in the music, which doesn't suggest anything particularly French. This quality seemed a perfect foil for the golden vaulted setting which I now envisaged. It would also highten the filigree quality of the enamelled fairy costumes. The courtiers' costumes were literally built.

Huge encrusted panniers were weighted with fur in order to accentuate the delicacy of the fairy costumes. At the outset, the director of the theatre was decidedly against the whole concept. He felt convinced that I was designing for a film studio, not a theatre and least of all for a ballet company. One of the major objections was that with such weight and elaboration in both sets and costumes – one would defeat the other, the result being indigestible to the eye. But I knew that the total effect would be a cohesive one in which the overall encrustation becomes a harmonising texture.

Then after all the battles to mount our production were over, on the first night, the director came up to me and generously conceded – 'You know we never thought it possible, but it works. It does work.'

This Sleeping Beauty was one of the largest theatre productions to have been seen in Germany since the war.

What about completely new ballets?

Yes, where you're working in unknown territory.

The choreographer has an idea to do a new ballet. What happens then?

Can I talk about Isadora which I am now designing? For this production we have been working from a scenario by Gillian Freeman. A very complete script and more than any outline I have worked with previously on a ballet.

Although it's unusual, the method is suited to such a narrative work. In some ways it is like a miniature film script, partly because of the numerous sequences. An idea had evolved for the sets which seemed workable and this would accommodate all the scene changes, so we decided that it was the way to continue and I developed a model to a state of completion. Then as MacMillan began to work on his choreography, something appeared to be too static about the concept. There wasn't sufficient flexibility for the fluidity which was now emerging in both choreography and production. After such a long period absorbing the original idea, we were both tentative about admitting the need for a re-think. But while MacMillan worked with the scenario, scenes were merging in a less realistic manner than anticipated, and this greater freedom reflected very much on the way in which I had designed the sets. So with a deep breath, we decided on a total re-design which I started only a couple of weeks ago. So you see such things happen of which an audience is quite unaware.

This kind of reconsideration is far less likely with costumes. There's the one example I've cited, but usually planning takes place over a long period and the costumes tend to come later not only because of the working method but the dancers' availability for fitting has also to be considered. For most ballets the making of the costumes is condensed into a shorter period than the rest of the production and often commences towards the latter part of its preparation. On almost every ballet production I have worked, the last four weeks is the crucial period for the costumes.


Three-act version, Act II

Inverse rendering of
preparatory costume drawimgs
for Tsarist relatives and guests
at a coming-out ball

The Royal Ballet
Royal Opera House
World Première 22 July 1971


THE SLEEPING BEAUTY, Prologue Set Design, drawing by Barry Kay, choreography Kenneth MacMillan, Ballet of the Deutsche Oper Berlin, 1967

Don Quixote

Costume designs for apparitions
Prologue, vision scene

The Australian Ballet
Her Majesty's Theatre
Première 28 March 1970
Adelaide Festival

Private collection, Canberra


Do you find that you tend to work with the same costume makers?

Obviously you develop certain favourites. That happens. Nowadays, not necessarily a great deal of work is done within the opera house itself. As you know, I work mostly for opera houses and in a certain number of these theatres like those in Vienna, Paris or Berlin, they do the greatest part in their workshops, but at Covent Garden it's not always possible. Quite often there are many outworkers, so you find some people who are better at interpreting your work than others and naturally I try to work with them if possible. The theatre fortunately understands that and makes it possible for those costumiers to be employed.

Presumably you are given a budget to work to?

Yes, but on certain productions this has not always been so clearly defined. Normally the managements have a reasonably clear idea of the average cost of a production, depending on its size and this is weighed up against the knowledge they have about a designer's previous work. Each department must also assess its own estimate of costs. The wardrobe supervisor will do this basically from the cast list. It is reasonably simple to gauge the average cost of a ballet costume which must include wigs, shoes and accessories. But with the sets it is a more complicated process because the individual parts are always different and the costing depends not only on the materials required for their making but also on their size. If special mechanical devices are part of the scheme, then these are also included in the total budget.

Should the designs appear to be going in excess of the budget, then the management will ask what you can do to help. This means looking at the drawings and the models and somehow you usually find ways around it. With experience you discover that it is not always necessary to stick to the letter. In the beginning I used to become easily disheartened over changes, but later you find that there is always an alternative to most problems. As Picasso said, 'when you don't have blue use red' or vice versa. It is like that, you know. Adapt to the means. The most important thing is the idea, that's what matters. Of course it's fine if you have all the freedom of an unlimited budget. But above all it's still the quality of ideas which is the essence.

24 October 1980, London



"Designing for the Dancer"
Authors: Roy Strong, Ivor Guest, Richard Buckle, Barry Kay, Liz da Costa
Foreword: Dame Alicia Markova
Publisher: Elron Press, London, 1981
ISBN: 0-904499-11-1
>> back to text


Although not related to the subject of this interview, namely ballet design, another mishap occurred when Barry Kay conceived the costumes for the opera Die Meistersinger von Nürnberg for the Royal Opera, Covent Garden. It was a mistake based on an obvious misjudgement that only became apparent at the general dress rehearsal the day before the opening night. As this opera features an immense cast, the stage was tightly packed with more or less motionless soloists and chorus, insomuch that Kay's superbly designed period costumes merged into one static mass, rendering them indistinguishable from each other. After weeks, even months of work, Kay was startled and – horrified. It was a tough lesson for him to learn that operatic acting tends to proceed in a rather stationary manner, whereas ballet dancing intrinsically requires continuous motion in which costumes can be seen individually.

In order to set the costumes off against each other, Kay applied bold black trimmings to accentuate their contours. At first, while the rehearsal was still in progress, a handful of garments were hurriedly prepared and brought on for the next act. It promised to be the solution. Due to the huge amount of costumes needing attention, enhancing all of them was certainly no easy task to accomplish. It would never have been possible without the concerted effort of the theatre's entire wardrobe staff working throughout the night. With no further rehearsal, it was touch-and-go all the way to the nerve-racking moment of curtain rise on the first night – but behold, it worked! >> back to text


arrow left    arrow up    arrow right