Bow Changes and Starts


When we “run out of bow” (reach the end of it) we have to “change it”. Fortunately for our economic situation, these terms “changes of bow” or “new bow/old bow” refer only to changes in the bow’s direction rather than buying a new one ………

If we were to use the bow as a weapon, paintbrush, back scratcher, fly swat, crowbar, cooking utensil etc, its range of possible movements would be enormous (and we would be obliged to actually replace it more frequently). If we use it however only for playing the cello, its range of movements is considerably reduced – especially while it is in contact with the string – and therefore, this “change of direction” to which we refer here is simply the alternation between the up bow and the down bow.


“Changes of Bow Direction” and “Bow Starts” are normally so intimately related that they are almost the same thing. Even in the most legato bowchange there is always a moment, when the direction changes, at which we have achieved a new (although almost imperceptible) “bowstart”. The exception to this “rule” occurs in portato, staccato and hooked bowings. Here, our new bowstarts follow on from each other in the same direction with no bowchange.

A new bowstroke (bowstart) can be smoothly connected to the bowstroke that preceded it, completely separated from it, or any degree of separation along the continuum between these two extremes. When, between two bowstrokes, there is a separation – a silence, a breath, a gap, a relaxation of pressure, a diminuendo – then that bow change becomes “easy” because the connection with the previous one doesn’t need to be maintained and we don’t need to disguise the new beginning. The greater the separation, the more irrelevant the previous bowstroke becomes and the more we can basically ignore the “Bow Change” and just consider each new stroke independently as a “New Start”. So let’s look first at bowstarts.



Clearly articulating a new bow stroke is the equivalent of tonguing for a wind instrument, or of starting a note with a hard consonant for a singer. For clean, crisp bow starts (articulations) – the opposite of legato smoothness – it is important that the bow be gripping the string before we start the note. Like that, the new bow starts actually with a sudden release of pressure, a mini-explosion. This is especially noticeable (and necessary) in dotted rhythms that we play with “hooked” bowings. Here, in order for the rhythm to be clear and precise, we need to grip the bow well into the string before sounding the short notes, as in the following examples:


If on the contrary, we wish to start our new bow stroke smoothly, we need to connect it perfectly, imperceptibly, to whatever came before, whether that be a previous note or simply a silence. Singers (and wind players) can teach us a lot about note starts. When they want to start a note imperceptibly from a silence (da niente) they start exhaling before starting to sound the note. For us string players this imperceptible note start requires several preparatory actions:

  • placing the bow on the string and then
  • starting it moving with infinitesimal pressure.

Both of these actions need to be done in anticipation of the true rhythmical start of the note. We need to banish the concepts of “off-on” or “yes-no” from our smooth note starts, replacing these with the more artistic and fluid concept of a light dimmer. We can start notes with the rhythmical accuracy and clarity of a gunshot – spiccato and very rhythmical passages often need that precision-  but in smooth lyrical music an exact start to the notes is often undesirable because it not only causes tension but also makes us sound like a computer.

In fact, placing the bow on the string before the note start is not only essential for soft gentle starts but for almost all note starts that are not ricochet or spiccato passages. Even a sffz attack will normally need to start with the bow on the string previously. This is another example of the absolutely basic Anticipation Principle.

Thus the technique for a seamless legato bowchange, and that for a beautiful smooth note-start after a silence, are similar but not identical. The legato bowchange is easier than a smooth bowstart in the sense that the new sound is not coming after a silence, and also because the body is already in motion, but on the other hand it’s harder than the new bowstart because the arm and body are moving in the opposite direction to which we want to go. Perhaps the secret to a beautiful bow start after a silence is, in fact, to treat it as a bow change: in other words to use as an anticipatory movement a sort of upbeat in the opposite bow direction. And perhaps the secret to both beautiful starts and changes is to anticipate them (start them slightly ahead of the beat).


How can we practice our bowstarts in the most efficient, concentrated way? It’s easy: with portato and staccato exercises, in one stroke (literally) we can do a whole lot of bowstarts.

While portato bowings work on our gentler, smoother starts, staccato bowings do the same for the more articulated, cleaner starts. For our crisp clean starts we can simply do the same portato exercise but now with staccato dots on each note. With staccato articulations we will be able to cram more notes (starts) into each bowstroke than in the portato because each note uses less bow in staccato. We can also have a faster note-speed with staccato than with portato.

But in both cases (portato and staccato) this is an ultra-efficient way to work on our bowstart skills as it not only crams many starts into each bowstroke but also does them in every different part of the bow and in both up and downbow directions.


During silences we normally take our bow off the string. This is for two main reasons:

  • if we were to leave it in contact with the string we would have to carefully control it to make sure it doesn’t make a sound
  • having to exercise this level of control eliminates the golden opportunity that the silence gives us for a moment of physical relaxation.

The fact that we take our bow off the string in silences means that smooth starts after a silence are normally “from the air”, requiring the placing of the bow on the string just before we start to “play” (sound) our starting note. This adds another degree of complication to our smooth-start process.

Violinists and violists are lucky because their instrument is attached only to them (and not to the floor). This allows them to accompany the placement of the bow onto the string with a cushioning, downward movement of the instrument. This is a beautiful, dance-like gesture which gives great control over the way the bow starts its stroke.  Unfortunately we cellists cannot imitate this gesture because our instrument is attached to the floor and has thus no possibility to move in such a way as to accompany (cushion) the placement of the bow on the strings. There are two ways to make up for this lack of cushioning, and thus limit the risk of bumpy landings (and crash landings):

  • keep the cello more vertical, so that the bow weight does not fall so directly onto the string. The more horizontal the cello (the longer the spike), the more difficult it becomes to control the bows landings (see Seating Posture)
  • turn the bow so that less hair is in contact with the string. When we do this, the bow hair is less tight, offers less resistant to the string and thus acts as a “shock-absorber” more than when we use the full hair. This is especially useful for delicate bow landings, changes and starts.

For more about the subject of placing the bow on the string from the air see Bow Trajectory in the Air.



Making the most of the movement possibilities of the wrist, knuckle and finger joints is vitally important for achieving beautiful bow changes, both legato and non-legato. How do we do this then? As we near the end of our upbow, if we let …….

  • the wrist come “up”
  • the knuckles go “down” and
  • the fingers straighten a little

……. then all this provides a sort of spring, a cushion, a “shock-absorber” to facilitate and smooth the change to the down-bow. By the time the bow has changed to the down-bow direction, the fingers have curved again, the knuckles have come “up” and the wrist has gone back “down”. And for the change from down-bow to up-bow we can do the exact reverse.

If we want to paint smoothly and ergonomically with a wide heavy house-painting paintbrush, we need to follow the same general principles at each change of brush direction. However the hand’s movements when painting with a large paintbrush are considerably simpler than bowing, as they are limited almost exclusively to the wrist and exploit more uni-directional movements than the bow-change movements (see below for a more elaborate description of the similarities between painting and bowing).

Normally the change from up-bow to down-bow is more problematic and less “natural” than the change from down to up-bow so we may feel the need to work more on the former (upbow to downbow change). One good way to do this is to play and observe “hiccup bowings” as in the following example:

wrist flexibility bach gigue 1Here, during the short upbeat (up-bow) we vigorously “throw” the bow from the string up into the air. The high energy of this “throw” not only exaggerates the different hand movements mentioned above (wrist, knuckles and fingers) making it very easy to observe them, but also in fact makes these movements doubly necessary.


In music we often need to connect the sounds in a smooth, fluid, singing, vocal manner (legato). In string playing this means that we need to be able to change the bow (direction) as imperceptibly as possible so that the musical line remains uninterrupted. In a resonant acoustic, the resonance makes the legato for us. In a dry acoustic however, these legato bow changes become more  difficult, and the closer we are to the frog, the more difficult they become.

By far the greatest part of this discussion will be dedicated to legato bow changes as these are the bow changes that cause us the greatest problems. Our legato bow change is the equivalent of the wind player’s circular breathing.  The constant, smooth alternation of opposite bow directions has its equivalents in life and the physical world. Before we start looking at actual bow-changes, a look at some of these parallels can be instructive in helping us to understand bow changes and how to achieve a smooth flowing reversal of bow direction.


In life, the closest equivalent to bowing is the breath. The down bow with its tendency to relaxation and diminuendo as we go towards the tip, is the equivalent of breathing out (expiration). The up bow, with its tendency to increasing weight and tension as we go towards the frog corresponds to breathing in (inspiration).

In the purely mechanical world, the equivalent of this bowing cycle would be the repetitive backwards-and-forwards movements of many machines (motors, pistons, pumps, sewing machines).

Where life meets the mechanical world (people using tools) we can see several good parallels such as the use of a paintbrush, a saw, and the alternating pushing and pulling of the accordionist etc.

Let’s look now in more detail at these different comparisons with the art of bowing.


Even though breathing is the closest musical equivalent to bowing, the legato of singers and wind/brass players cannot in fact teach us much about bow changing. This is because they only use one side of the cycle (the expiration) to make their sound. One expiration for them corresponds to a single bow stroke for us. It is as though they only ever make a sound with down bows (very slow ones) while the up bows (silent and rapid) are just to breathe!

So whereas for string players the main impediment to a flowing smooth legato is the frequent changes of bow direction, for singers, the impediment to legato is the occasional need for a new breath. Or, put in other words, whereas we string players must change bow direction imperceptibly and frequently, singers just stop every once in a while, take another breath, then start again.

Playing only with down bows would pose a serious legato problem for us. But for singers, the situation is saved by the fact that their breaths are much longer than our bows. Until the invention of a retractable 3m bionic arm, the necessary shortness of our bow is such that one bow stroke can only sustain the string vibration for a fraction of the time that a singer can sing on one breath. During that one breath (expiration) it is no easier to make a beautiful perfect seamless legato than it is for us string players during one bow stroke (and on one string) – the breath just happens to last soooooo much longer.

We can however learn a huge amount from the circular breathing of wind and brass players. Here, while part of the respiratory system (the mouth) is blowing air out, the rest is sucking air in (through the nose). This illustrates a vital concept for legato bow changes: we actually NEED to be able to move in opposite directions at once. In other words, we need to keep the bow moving in one direction while rest of the body has already turned around and is going in new direction. It sounds like a magic trick, an impossible Exorcist-style (film) special effect, but the following section, with the help of paintbrushes and reptiles, will show how this “sleight of hand” is possible.


Take a piece of rope and lay it on the floor. Pick up one end and move it (wiggle it) from side to side. If the rope is the right length (not too short and not too long), and if you move it at the right speed and amplitude, the loose end of the rope will also move on the floor from side to side. Its movements will copy the movements of your hand exactly ……….but with a delay. So when our hand is moving to the right it is quite possible that the far end of the rope is moving to the left (or vice-versa).  This delay is because the rope is flexible. If we try the same experiment with a long stick (and a stiff wrist) there will be no delay in the transmission of the changes of our hand direction to the tip of the stick (because there is no flexibility).

The same phenomenon can be seen in the movements of real live snakes ……. and in a human snake (a long line of people, each holding onto the person in front of them and all following the leader).

Looking at how we use a paintbrush is another instructive example and is even more useful to us string players than the rope example, because the paintbrush, compared to that piece of rope we just used, has a lot more in common with the cello bow.

Even though our brush strokes follow the same perfectly straight line in both directions (exactly like our bow strokes), the moment of change of direction is not simultaneous along the arm-wrist-hand-fingers-brush handle-bristles chain of command. Firstly the arm changes direction, then the wrist, then the knuckles, then the fingers, then the brush handle, and only then, the bristles of the brush. As with the rope, it is the enormous flexibility of the bristles that allows us to see this process of “delayed reaction” very clearly. It is this “delayed reaction” that gives smoothness and fluidity to the change of direction.

Unfortunately the bow is not quite as helpful as the paintbrush. Whereas the bow stick and hair do have flexibility, unfortunately this flexibility is not in any of the directions that could help us with our bow changing. This means that all that flexibility necessary for a beautiful, smooth, paintbrush style bow change must come from our own body. It is our fingers, knuckles and wrist that now have to mimic the flexibility of the paintbrush bristles.


Have you seen videos of oil well pumps, piston engines or other back-and-forth-repetitive-movement machines? The way these machines are designed can teach us two additional important lessons about how to reverse direction smoothly, even for machines with solid, hard, rigid moving parts in which there is no flexibility available to ensure the delayed “snake” reaction that we saw with the rope and the paintbrush.

Let’s not waste time: the moving parts that drive the simple backwards and forwards mechanical motions always:

  • move continuously and
  • describe elliptical shapes.

These then are the two lessons, except that they are really only one lesson: that the secret to all smooth 180º directional changes is that there must never be an absolute stop to the movement.

The way to ensure this is, in the absence of flexibility, to convert the moment of change into an elliptical, circular movement. Without this circular component to the change of direction, at some point, just before the change, the movement has to stop completely. This is bad! To stop completely, you have to slow down – this is also bad (but not as bad as stopping dead). Basically, all slowing down and stopping to change direction is an interruption, both musical and mechanical.

Please note however that in most machines, the piston (or “working” part, corresponding to the bow) follows an absolutely straight line (with no elliptical movements at the direction change) and does therefore have a moment at the change of direction when it stops moving. However, in these cases, it is the parts that are pushing and pulling this “piston” (corresponding to the arm, wrist and hand in the case of bowing) that follow elliptical paths and never stop moving (see oil well pump video). The bow, unlike a piston, is however not obliged to follow a perfectly rectilinear path at the bow change. Thus we can make the movement of the bow at the bow change slightly elliptical, in order to keep the bow moving and avoid a “dead” moment at the bow change.


For a seamless cello legato bow change:

1: the arm must never stop moving
2: this requires that it move like a paintbrush (snake)
3: we can add an optional elliptical movement at the bow change.


At the tip of the bow, we use the bow in exactly the same way that artists uses their paintbrush: the forces and movements applied by the hand are transmitted directly to the point of contact of the brush with the canvas (or the bow to the string).

At the frog, the situation is more complex. This is due to the lever effect – or to be more exact, the fulcrum effect- when playing at the frog. It is as though the artists brush was long and heavy, but the actual hairs of the brush were now protruding from the stick from a point very close to the artists hand. To understand this fulcrum effect for the cello bow, imagine now that your is bow very short – the length of a pencil for example. When the bow is very short, there is no difficulty changing at the frog, and there is also no difference between changing at “the frog” and at “the tip” because there basically is no tip –  it’s all “frog”! Imagine now that your bow is enormously long. Now, changing at the frog becomes much more difficult. This is because there is now so much more bow weight situated on the other side of the fulcrum (point of contact with the string). The longer the bow, the greater this weight becomes.

While playing at the frog, all this extra weight has to be controlled. But it is not only a question of weight. The closer we are to the frog, the more the forces produced by this weight are amplified by the lever/fulcrum effect. To illustrate this phenomenon, make a small string crossing movement at the frog and observe how the movement is at the tip of the bow is so much larger.

It is this combination of extra weight to manage + the fulcrum effect, that makes it more difficult to control the bow changes at the frog than at the tip …… and why artists paintbrushes always have the hairs at the tip of the brush!

This is also why the baroque bow: shorter, lighter, with the balance point basically AT the frog, is so much easier to manage than the longer, heavier “romantic” bow. The baroque bow is a more delicate brush especially when it is held with a “baroque” bowhold (in which the hand holds the bow at a position slightly more towards the balance point).