Choosing Bowings for Asymmetrical Figures

Normally we plan our bowings to avoid sudden and extreme changes in bow speed (unless deliberately needed for special effects). This planning is especially necessary in figures in which the alternation of long and short notes note creates “asymmetrical” rhythms (see Dotted Rhythms). In asymmetrical rhythmic figures we basically have three possible bowing choices by which we can avoid unwanted accents: the retake, the hook, and the sleight of hand. Let’s look at these three possibilities now:


If we can lift our bow off the string before or after the short note, then this enables us to move our bow, silently in the air, back to where we want it to be. This, along with “hooked bowings” (see below) is one of the easiest way to solve the problems of bow division posed by repeated dotted figures. We have two choices for when to retake the bow:

1.1:  In a brief “gap” after the long note. This gap may be written as rest.

But at other times there may be no rest indicated between the long and short bows of an asymmetrical figure. It is surprising just how often we can make this “gap” after the long note, even when it is not specifically written as a rest. In these cases we will need to “steal” the time for the retake from the long note by shortening it slightly:

If, near the end of the long note, we release the bow from the string while still maintaining plenty of bowspeed (see Note-Endings), that long note will keep sounding even without any bow contact, and this resonance – like the piano’s sustain pedal – creates the aural illusion that we are still sounding it with the bow. During this resonance we will bring the bow back towards the frog. No music scores show any sign for this choreography, both because no standard sign exists and because there are other bowing alternatives to the retake (namely the “hook” – see below). This can be a shame, especially for inexperienced players. In all of the edited sheet music available on this website, rather than overloading the page with thick invasive arrows, the comma has been used to indicate this type of retake. This is however definitely not the optimum form of notating a retake because in most printed music a comma indicates a rhythmical delay.

In the above examples we are “retaking” to an upbow. We can of course also go “the whole way”, and retake to another downbow, as in the following example:

1.2: Sometimes we will, by choice or obligation, do our retake in a brief gap after the short note (rather than the situation in all the above examples where we retook after the long note).

The following link opens a page of repertoire examples showing the use of the retake in  many different circumstances: The Retake: Repertoire Examples


In those asymmetrical passages in which the musical articulation does not allow us to lift our bow off the string and thus “retake” it, we can avoid the bunnyhop accents (see below) by using “hooked bowings”. Here, we “hook in” the little note in the same bow direction as the previous longer note.

hooked bowings

The hook is an alternative to the retake. Here below are the same three examples that we used to illustrate the retake, now bowed with the hooked bowing instead:

If we don’t use the “hooked bowing” (or the retake) there are two possible (but not obligatory/unavoidable) negative outcomes:

  1. an accent is caused on the little note due to the fact that we need to use a suddenly much faster bow on it in order to get back to the end of the bow for the start of the next long note, or
  2. if we try to play the passage without accents on the short notes, our bow finds itself gradually working its way more and more to one end, which is why we could consider these bowings as “unsustainable”.

non-hooked bowings

The back legs of rabbits are much more powerful than their front legs. This is why their running rhythm resembles more a dotted rhythm than the regular binary flow of other animals whose front and back legs are more symmetrical. Playing dotted rhythms with hooked bowings or retakes avoids that sudden fast bow stroke giving this same “bunnyhop accent” to the short notes.


Unfortunately, not all asymmetrical figures are suitable for hooked bowings or retakes. Retakes require “air-time” to get the bow back towards the frog. Hooked bowings require that we can relax the bow pressure (and speed?) on the long note in order to be able to “restart” the bow on the short note. But sometimes we can do neither! In the examples shown below, even though the bow is playing dotted rhythms (as shown in the second line), the “music” is not “dotted”: the notes flow uninterruptedly. Because of this, the “hooked” bowings that can make our life so much easier when playing  dotted rhythms are much less appropriate here: we just don’t have time to stop and restart the bow (nor of course to retake it). So what then what can we do in the following figures to avoid the bunnyhop?


In these cases we need to make a choice between:

  • finding the best bowing with which to play the articulation as written. Even with absolutely the best choice of bow directions, we may still be required to be a little like a magician, using radical bowspeed (and pressure) changes to avoid the bow being taken to where the laws of physics want it to go rather than where we want it to be, but at the same time making those radical changes as imperceptible as possible to the listener
  • changing the articulation, while trying as much as possible to maintain the composers desired intentions

Our choice will depend on our answers to the following questions:

  • does the composer understand string bowings and deliberately wants to obtain the special “bunnyhop effect” (agogic accent) of the huge and sudden variations in bow speeds? In this case we might not change the articulation.
  • or, is the composer unaware that the articulation they are asking for, while perfectly uncomplicated for non-bowed instruments, is very awkward for a string instrument? In this case we could change it to something that sounds better, rather than sounding bad with the composer’s original suggestion.

Certainly, when transcribing music from other non-bowed instruments we have every reason to change the slurs to make the most asymmetric articulations more “bow-friendly” because it is only for bowed instruments that these asymmetrical articulations pose a problem. The following passage is originally for flute:


Often the “reverse bowing”  (upbow on the beat, and short note on a down bow) is more comfortable (as in the above example). This is especially so when the short note is on the higher string (as above) because the string crossing is greatly facilitated by going in the “right” direction (see String Crossings). But even if the short notes were to be on the lower string this “reverse bowing” still has many mechanical advantages.

Sometimes we are lucky, and the asymmetries counterbalance each other in such a way that we don’t need to change the bowing nor take any radical bowspeed measures to keep the bow where we want it:


It’s like a very simple maths problem: we add up the beats going in one direction and then add up the beats going in the other direction ….. and then compare the two totals. For example, in the first four bars of the above example, even though the figures are all very asymmetrical, we do 7 semiquavers towards the tip and 5 towards the frog in each bar. This is quite balanced: this small difference is easy to compensate for with a minimal bowspeed correction.

At other times we can use the asymmetries to bring us where we want to go in the bow. If the phrase is in crescendo then we will bow it in such a way that the asymmetries take us from the tip towards the frog as in the following example from the same Bach Courante:


If the phrase is in diminuendo,we could bow it in the opposite way.


Here is a compilation of rhythmically asmmetrical repertoire excerpts for which neither the retake nor the hook can be used to alleviate the problems of bowdivision:

Rhythmically Asymmetrical Repertoire Excerpts