How to add a rest above a note?

• Apr 9, 2021 - 00:38

I'm trying to put a rest above a note, but Musescore keeps thinking that I want a rest to replace that note. Could any of you help me with this?

Note: Im not looking to use SATB for this. I'm using piano.


You can select the note, hit "z" to show up the master palette. Under "symbols" search for rests and click on one to insert it at your note. Then drag it to desired spot.

In that respect there's no difference betwen Piano and (Cloded Score) SATB: you need voices. Here apparently notes in voice 2 and a rest in voice 1

There has recently been an extended discsussion about this. Indeed the solution is to use Voices, but there are several elephant traps to fall into. Some of this (IMO) is due to user's mental models not matching what MuseScore actually does, but there are also some glitchy features which may not work well for every user.

I don't know why you want to put a rest above a note, but presumably you are hoping to fill out the rest of a bar with other notes - which may - though not necessarily - not have the same rhythm as the other notes on the same stave.

I attach a very small file containing a not overly musical fragment of two bars. You'll notice that there is a rest over the first note. Also note that there are effectively two parts, and they have different rhythms.

Additionally If you select the bars you'll notice that the colours of the notes change. The blue notes are not very distinguishable from the previous black ones, but the green ones stand out more. There is another PDF file containing images attached to show this. The blue ones are - in MS terms - Voice 1, while the green ones are considered to be Voice 2.

You will also notice colours if you switch into Note Input mode - with shortcut code N. This may, or may not, be helpful and/or confusing to you - but at least it might give you a clue as to which input mode you are operating in.

As it says in the handbook, don't necessarily think of the voices as being associated with SATB voices.

It is possible to swap over the voices : Tools -> Voices -> Exchange Voice 1-2 and the result of that on the example is also shown as an image in the PDF file.

You'll notice that the green notes have flags which go down, the blue ones have flags which go up.

If you want the flags the other way round, you can change them by selecting one or more notes (bars) and using the x command.

It is also possible to select all the notes in a stave in the same (MS) Voice - and apply operations to them as a group - for example reverse the flag direction of all of them.

To do this select the first note in a voice (it doesn't seem to work if a rest is selected) - hold down CTRL while doing that. A sequence of pop up menus appears - first find Select - then another popup appears - at the bottom of that one there is More - click on that - then check the box for Same Voice. You might even be able to apply or more characteristics - e.g. Same Bar or Same System at that stage. That is really clunky. Then click OK.

Then apply the operation - for example x to reverse the flag direction. Other operations, such as raising or lowering notes can also be applied, or even transposition. It might be difficult to discover whether the operation is going to be applied to just one bar, or all the remaining voice notes in that stave.

Doing the select is quite tricky - as mentioned about - can be clunky.

Note also that you can have 4 voices on each stave. The colour codes appear to be 1 Blue, 2 Green, 3 Red, 4 Purple - but the purple and the red are really rather similar - which isn't a great feature, and I've already mentioned that the Blue is not so obviously different from the black. This would fail an accessibility test - such as - particularly the red/purple colours. D8008C vs D92F00 giving a result of 1.01 = FAIL.

For the purposes of generating printed text you might find it convenient sometimes to use the extra voices 3 and 4. For example, if you only wanted to modify the voice parts in a few bars, say make changes to voice 2, you could exchange voice 2 with voice 4 for those bars, then make the changes. You could even leave those changed bars as voice 4, as when printed the notes would all be black and I don't think they'd be distinguisable from voice 2 notes.

Good luck with this - I hope the examples and the images help. Experimentation might be the way - I fear - until some significant changes are made in this feature of Musescore.

I put several different examples of the colouring/voice assignment into a PDF file - also attached.

Attachment Size
test-voices.mscz 4.47 KB
MS Voices example.pdf 112.61 KB

In reply to by dave2020X

Voices indeed have got nothing to do with SATB voices, AKA human singers and their voice, but has to do with Closed Score SATB, where 2 voices (Soprano+Alto resp. Tenor+Bass) are notated in one staff. Same for a single staff for e.g. Flute 1 and 2 or Violin 1 and 2, in general for more than one, usually monophonic, instrument sharing a staff. And then even if they are using the very same rhythm, so there wouldn't be a real need for using voices.

In reply to by Jojo-Schmitz

I'm not sure I'd go so far as to say nothing to do with SATB voices - but essentially voices in MuseScore are a device for layout, printing and engraving. It's even more confusing if some voices are actually playing chords - not possible with human voices - except perhaps for some throat singers, but for instruments that may make sense. The layout of SATB human voice scores can be implemented - arguably more easily - by the use of MuseScore's Voices - but it can also be very confusing.

In that sense MS's voices are not the same as voices in a vocal score, and as I indicated in the response, if the notes are all printed off using black note heads, there would be no obvious difference. MS voices are a device to get the kind of layout which musicians might want in a printed text.

There might be an issue if separate parts for each singer are required - to make sure that the correct MS Voice is assigned to each singer - though normally that would not be a problem. I suppose there could also be an issue if different vocal parts required different lyrics - which I guess must happen sometimes. These issues are probably most often sorted out by people with more experience of the printing/engraving conventions, and of the particular notation package being used for generating and editing the scores.

In reply to by dave2020X


Voices in a guitar score are not about human singers at all but are a means of having independent rhythms. A 6-string guitar could theoretically need a 6-voice stave but in the real world 2 or 3 voices are enough, e.g. a melody tune + accompanying bass notes would use 2 voices.

Specialised Guitar TAB software also uses the term "voices" for these independent rhythms.

In reply to by Jojo-Schmitz

Have we heard from the OP why he/she wanted this, and whether the question has now been answered?

Unfortunately to use a term from Martin Keary - this is one aspect of MuseScore which I feel new and even some not so new users might find Janky - but I do agree that it is necessary to use voices.

In reply to by dave2020X

To be clear: "voices" in music - and hence in MuseScore - are a fundamental feature of standard music notation. So yes, it's all about layout, printing, and engraving, just as with almost every feature in MuseScore. Voices in MuseScore mean precisely what they mean in other musical contexts as well. Which is to say, an independent melody line. It's the same as Bach means if he write a fugue for organ but says it is in four voices. It doesn't mean he wants four people to sing it, but rather, that there are four independent melodic lines.
Whether the voice happens to actually be sung is immaterial - the sound used has nothing to do with the usual definition of what a "voice" is in this context. It's about one thing and one thing only - multiple independent lines. In this case, specifically, within a single staff.

In reply to by dave2020X

It's not exactly a rigorous definition, true - and it's something not really that commonly taught to beginners reading music, which is why people who didn't specifically study composition may not know the term. An unfortunate situation to be sure, but not one MuseScore can really do anything about. We need to use the correct terms and correct mental model and correct notation, but educating people on what these are and how it all works takes effort. It's a lacking of the music education system, I would say, that people can otherwise learn to read music but not be well-versed in this particular aspect of it.

As for "chords and other combinations of sounds", that isn't really any special extension - it's centuries old as well. Again, if you stop thinking of voice as necessarily meaning a single human but instead the idea being wrapped up in the idea of independence - specifically, of the rhythm - you realize that it's completely possible and in fact quite common for a single voice to actually contain chords.

In reply to by Marc Sabatella

I don't have a problem with a voice being a sequence of sounds with rhythmic independence, including chords, but trying to find a strict definition appears to fail.

You mention using the "correct" terms, but who decides that? Would you consider an extra "tape" track with birdsong etc. to be yet another voice? It would not necessarily fit into any notational context rhythmically, but composers have used such techniques.

My hunch is that most people, including many musicians, have an intuitive understanding, but that it's not quite as precise or universally well understood as you imagine.

In reply to by dave2020X

I think that an extra track would have its own stave.

The definition of voices is consistent in the 3 scoring programs that I have used: MuseScore, Guitar Pro and TablEdit so it may be a standard term.

See also this score from BSG where he has taken 4 separate instrument parts and made a 4-voice piano part from them.

I'm sure that "The Bachmeister" himself is using the correct musical term.

In reply to by yonah_ag

When you write "The definition of voices is consistent in the 3 scoring programs that I have used: MuseScore, Guitar Pro and TablEdit so it may be a standard term." you are really declaring that your understanding of the methods and techniques of the three programs seems, as far as you can tell, to be the same. That does not mean that the definition of voices in the programmes is necessarily identical.

I have looked for a formal definition in multiple sources, and not found one. That doesn't, of course, mean there isn't one - but all my attempts find a formal definition based on established music theory or writings on music which is consistent with what has been discussed have failed. Sure - most of us have an intuitive notion of what that means. I am happy enough to go on using the terminology in a loose way - and to observe the way that MuseScore implements what it calls "voices", and how the voices in MuseScore are used by people trying to write or use music notation.

To summarise so far:

  1. MS Voices are not necessarily associated with human voices.
  2. MS Voices allow strands or sequences of notes and rests with different rhythms to run in parallel.
  3. MS Voices may include chords.

Also - currently MS Voices do not necessarily have an obvious effect on the printed notation for notes, as all the notes are printed in black. There may be an effect in that the stems of some notes may change direction, though that can be achieved by other means. Generally in a stave Voice 1 and Voice 3 - if not modified - will result in the same appearance - usually with note stems going up. SImilarly Voice 2 and Voice 4 - if not modified - will result in the same appearance - usually with note stems going down.

During development of a printed score, the visual appearance of notes and rests in MS Voices on computer displays may change, and become coloured, but this does not affect the final outcome. At the present time there is - AFAIK - no other significance for MS voices than to generate the printed layout. This could change in the future, though I have no indication that it will.

In reply to by dave2020X

"currently MS Voices do not necessarily have an obvious effect on the printed notation for notes, as all the notes are printed in black"

MS Voices do show up in different colours during Note Input and whenever you select a note.
MS Voices also show up in different colours if you do a Range Select. This is very useful as a proofreading tool.

"the visual appearance of notes and rests in MS Voices on computer displays may change, and become coloured"

See the scenarios above. The other case in which notes (only notes, not rests) become coloured is when the specific note is outside the range of an amateur (orange) or professional (red) performer. This has nothing to do with Voices, and is purely about the pitch of the note.

In reply to by dave2020X

@dave2020 : When you write "The definition of voices is consistent in the 3 scoring programs that I have used: MuseScore, Guitar Pro and TablEdit so it may be a standard term." you are really declaring that your understanding of the methods and techniques of the three programs seems, as far as you can tell, to be the same. That does not mean that the definition of voices in the programmes is necessarily identical.

I can only speak from experience as my music theory knowledge is limited. Other comments in this thread also support the same interpretation of "voices" and you have provided no evidence that another term should be preferred or that the term "voices" is not correct.

"As far as you can tell" ????

No, it's not just as far as I can tell, it's also in the handbooks for these products.

In reply to by yonah_ag

The term "voice" and voicing has been used for centuries in Music Theory.
One such example is even at the heart of the definition of a Fugue wiki:
[…]In music, a fugue (/fjuːɡ/) is a contrapuntal compositional technique in two or more voices, built on a subject (a musical theme)[…]

Within MuseScore, notes in a separate voice can even be assigned their own channel through a staff text and thus use a different playback sound.

In reply to by jeetee

I am not disputing that the term voice has been used for centuries in the context of contrapuntal music.
It's when the "voice" itself contains chords that I start to wonder if that was what was meant, centuries ago.

The verb form "voicing" has also been used for a long while in the context of instruments, such as organs.

Re the MS use of voices - do people actually use the different channel assignments for different voices?

I do accept that this may be possible in MS, and I'm probably not even going to query whether that might be useful. I personally don't know how to do that yet - others might. I attach two files. One is .mscz with two very boring fragments apparently in a couple of voices - if printed in black and white. However if they are shown with the colours then the second iteration shows that actually four (4) MS "voices" are used. A musician just looking at the black and white score would say there are two voices in each stave. The PDF file show the colour assignments, indicating that these are different MS voices.

I also show another file with chords as well as single notes in the "voice" parts. Was it always accepted that chords could be part of a "voice" in music theory - or did it just evolve - as I suspect?

Attachment Size
test_voices.mscz 8.49 KB
test voices.pdf 197.52 KB
test_voices2.mscz 10.17 KB

In reply to by dave2020X

This depends on the place of use.

See for example the accompaniment-part below:

Now look at this:

Both have equivalent chords.

The difference between them:
* The first one is accompanied by an arpeggio,
* The second one consists of only chords.

So, just like the first one, it doesn't require a separate voice/voice-line. (Or we can combine as many notes as we want with the same note value (duration) in the same voice/voice-line and use it as a chord)

Now let's create a technical problem and replace the tied notes in the 3rd measure (maybe just because they look prettier or more legible) with the whole note value (duration):

Now there is technically necessary to use two different voices/voice-lines. But actually these are two separate chords: C and C7 and it could be written in a single voice/voice-line, as in the previous example. This is how we see it because we combine common notes in one value (duration) instead of tied them.

In reply to by Ziya Mete Demircan

Thanks - that's interesting. Do we know when composers became really aware of this kind of technicality?
If we assume a model - top line = melody, bottom line = harmony - though there may have been other models, then this kind of issue might have been written about in, or before Bach's time. It may even have been the subject of discussion way back - say in Palestrina's time, when there would have been other issues about associating words with music.

In reply to by frfancha

I didn't suggest that ZIya's comment wasn't informative. Indeed I think it is, but it does not allow us to determine when chords became treated as parts of "voices". It is possible that it was in Bach's time or in the later classical period, or much earlier - even in the period of polyphony. Without some form of evidence - probably written - by music theorists or commentators at the times in question, it would be hard to know what the common practice was, and what terminology was used.

Using examples is an ostentive way towards a definition, but does not provide a formal definition of techniques and terms which were used two, three, even five or six centuries ago. I'm not objecting to the current way in which voices are defined in MuseScore, but probing the claim that this is exactly the way they have been defined and used for many centuries.

In reply to by dave2020X

The concept of multiple voices is absolutely universal, known to anyone trained in music theory for the past several centuries. But of course, computers weren't a thing in Bach's day, or Mozart's. So the specific ways in which any given computer program might happen to implement the controls for multiple voices won't be anything Bach or Mozart would necessarily recognize. And just as different programs provide different basic note input commands even though they all necessarily have the same underlying notion of what a note is, so can they provide different commands for working with voices even though they have the same underlying notion of what voices are.

So instead of trying to figure out the One True Precise Definition Of Multiple Voices And Model For How They Must Work In Any Computer Program - spoiler alert, it doesn't exist - best to just understand the basic general concept that is common to every single program, the same basic knowledge common to every single person who studies music theory. And then you can hopefully wrap your brain about the specific ways MuseScore happens to implement this universal concept.

In reply to by dave2020X

Since music in most cultures began as completely monophonic, anything involving multiple pitches notes sounding at the same time necessarily evolved gradually. But consider a voice as the notational concept we are talking about here, most likely people started notating chords within a single voice very shortly after first hitting on the idea of putting multiple noteheads on a single stem. Not sure when that would have been, but again, surely many many centuries ago.

Anyhow, it's an interesting thought experiment to consider possible derivations of all this, but I would also say, if the goal is simply to understand how to use multiple voices in MuseScore or any similar computer program, it's not really important to know whether the first instance of this was 1200 AD or 1400 AD. It is important to know that that it's an integral part of modern notation, and to understand when and how to use it.

In reply to by dave2020X


1. the sound which issues from the mouth, and which is produced by the vibrations of the vocal chords
2. individual line(s) or part(s) in a musical structure
3. distinct instrumental or vocal part(s) in a musical work, for example, a four-part song may be described as being for four voices, even if each line is performed by a more than one person, as in a large choir. see 'alto', 'soprano', 'mezzosoprano', 'contralto', 'tenor', 'baritone', 'bass'
(These were written in separate lines, I added the list numbers.)

Voiceline: an individual melodic line formed by a voice within one measure. When a staff contains more than one voiceline in any given measure, note stems of each voiceline usually point in the opposite directions. The sum of all note and rest values of each voiceline in any given measure should be accounted for, but sometimes are not.

In reply to by Jojo-Schmitz

Thanks. Interesting that it has to be hosted on - or does it work for other sites too?

I'm not sure how Ziya Mete got his image into the thread too. Since we now think this site uses BBCode let's try this [img][/img]
View post on

Nope - none of the "obvious" markup languages seem to work.

The music fragments appear to be hosted on - for example at… and the page here has been created using Drupal -maybe!

Looks like there's some markup which is not too widely known about. I've still not figured out how to embed an image hosted on and indeed whether the markup would work for images hosted on other sites.

In reply to by dave2020X

This site isn't using a lot of generic bb codes, it's using a limited version of MarkDown with some very limit bb-coded (and xmlcoded) extentions
To get an image here inline, you first upload it as an attachment to your post, then click the "insert" button/link in the uploaded line of that attachment.

There are no other sites besides that turn mscz into online playable scores; so no, embedding scores from other sites in such a way won't work.

In reply to by jeetee

This is just a test to see if I can post an image file with a simple demonstration of the colour coding of voices in MuseScore using the approach mentioned by jetee.

Yes!! It does work - but the name of the image file has to be surrounded by square brackets [ ] at the point where it is intended to be displayed. Also the green check mark also has to be selected/activated.

test voices.jpg

The colour coding does show up in bars which are selected - or indeed in the whole score if CMD-A is used to select everything.

Obviously it is also possible to - inadvertently - get two copies of an image displayed. Control over the display seems somewhat limited. It may be necessary to remove the uploaded file and reinput it to get the display to function as intended.

Phew - hopefully got only one example shown now.

Attachment Size
test voices.jpg 202.47 KB

ok, uh...
i see there is an argument in the comments
im trying to put a rest inbetween treble and bass, not directly above a note.
i'll post a picture of what im trying to do.

In reply to by Rizzieri

Thanks. Somwhat fuzzy - nevermind. But I found myself thinking "why would anyone want a rest in the place you show?" - and of course the answer emerges from the discussion here. The rest and the way the note stems are aligned does indicate that the bass staff is composed of two voices - and I'm not going to concern myself now with arguing about those.

It looks as though the left hand fingering is between 3 and the thumb, and the timing follows from the voices.
I wonder if there is always a rest at the start of bars where more voices are introduced - someone might comment.

Is it "bad form" to start a new voice part way through a bar - though that would avoid the need to put in the preceding rests?

Anway, I guess that you should now know how to do it in MuseScore.

In reply to by Rizzieri

The quarter is there to tell you that the "D" in the the top voice of that measure starts on beat 2,. Otherwise you'd need to guess by seeing how it lines up with other notes, and often there is no other note to line up with. That's why it is normally important for each voice to be read left to right with all rests shown explicitly. So the top voice goes quarter rest, half note; meanwhile simultaneously the bottom voice goes half note, quarter note.

In reply to by dave2020X

In general, as noted elsewhere here, voices should be complete, so each voice can be read left to right with no possible ambiguity about which beat any given note occurs on. So yes, you show all leading rests if a voice enters mid-measure.

There are exceptions, however. Different editors might have different rules they follow for the the corner cases, but in general, it's OK to omit leading rests if both of the following are true:

1) this particular voice is not persistent - it's not like a voice in a fugue that is expected to "live" throughout the piece, but rather a voice needed only to express a particular variation in rhythm in this measure specifically


2) there exists another note on the same beat as the first note of the voice in question, on the same staff. That is, you can look at that staff only and see exactly what beat the first note of the voice lines up with.

In this particular example, 1) is true, but not 2). The piece is mostly two voices only - one lives on the top staff, the other on the bottom staff. This is the only place (this and the following measure) where it splits into three voices, so the only measure where a second voice is needed on the bottom staff to handle it. But, in both of the measures, the top voice (the one with the rest) enters on beat 2, and there is no other note one beat 2 of this staff to align it with. There does happen to be a note on beat two on the other staff, so it's not like it would be impossible to figure out which beat the D comes in on, but that's not really good enough. Showing the leading rest is necessary to clarify the rhythm completely within this staff.

Hello everyone! I'll be closing this post now- seems like the only way to put a rest above a note is to use voices. Thanks for your help!

Oh wait nevermind i can't close it

In reply to by Rizzieri

Your piece is fairly obviously "Bach"'s Minuet in G. See bar 25.
Now attributed apparently to Christian Petzold.

Attached is my quick rendition of that/your bar, made using voices.

All that remains for me to do is to try to figure out how to get different sounds for Voice 2 compared with Voice 4.

You can see here Voices-A.png that when the staff or page is selected if the notes are in Voice 2 they appear in green, while here Voices-B.png the notes are input in Voice 4 and they appear in a kind of purple.

When finally rendered as black on white, both the Voice 2 and the Voice 4 versions will look the same.
However switching to Voice 3 - if you try that - will alter the vertical layout and the direction of the note stems.

In reply to by jeetee

However, I note this from the handbook - "Most of the time is it far easier to just apply the corresponding text (pizz., tremolo, arco, mute, open, etc) from the Text Palette."

I suppose that anything else, such as turning a voice sound from a violin to a clarinet or ukelele would really push this into the realms of Midi control.

Maybe there could be some use for this - for example if it's possible to mute one of the voices. A flute part written out with two flutes on one stave might have the two flutes using Voice 1 and Voice 3 for the 1st flute, and Voice 2 and Voice 4 for the 2nd flute. Then on playback there could be different combinations actually producing the sound.

Also, apart from muting, or getting different types of sound - such as arco, staccato etc for string instruments, can voices also be used to change the balance or volume of the sounds of each voice individually - say in the Mixer?

Whether it's really worth that I'm not sure. That might be the time at which one might decide to export the Midi version - or one of the other digital formats - to another tool - such as a DAW.

In reply to by dave2020X

Pretty much "yes" to all of the above.

If an instrument has multiple channels, then you can expand that instrument track in the mixer to show the subchannels and assign whichever sound you want to it; regardless of the name of the channel (which is only that, a name).
So yes, you can use such an instrument to assign completely different sounds to a voice (check for example the men/women instruments for a closed SATB score) via such staff texts.

Completely separate from the sounds assigned to an instrument and whether or not it has multiple channels, you can always mute certain voices for an instrument in the mixer as well, by pressing the relevant buttons in the mixer:

As mentioned before; yes you can set relative volumes for channels within an instrument by expanding the mixer track. So if you assign a voice to a channel, then you are indeed setting the voice relative volume.

In reply to by jeetee

Here is a simple two bar example - based on flutes.

It is possible to mute each of the voices (flutes) independently, but the mixer only appears to have one copy of the instrument, so it's not possible to adjust the relative levels of the parts.

I suspect that this depends on whether an "instrument" in MS is composed of different timbres, playing styles etc. If an instrument has sub-channels to generate the sound then it may be possible to adjust the relative levels, but not, it seems for flutes.

If one wanted to adjust levels independently then I think the instrument parts (for flutes, at least) need to be written on separate staves.

In reply to by dave2020X

The default Flute instrument indeed only has 1 channel to control (it is after all intended for a single flute).

This is where a custom instruments.xml (or manually adding a channel in the mscx) comes into play. I've duplicated the existing channel and called them "Flute 1" and "Flute 2". Now that you have separate channels, you can use a staff text to assign voices to them (as done in this example) and thus assign different mixer sounds to it.

In reply to by Jojo-Schmitz

So, a nice workaround suggests itself if you don't want to mess with custom XML files. Just add the instrument as "Women" or "Men" (which have the same two-channel design) and change the playback sound, staff name, and other staff properties as desired. it will then work as a flute but you'll have the same ability to work with the two channels.

For MuseScore 4 it would be nice of course if this kind of kludgery weren't necessary.

In reply to by Rizzieri

"seems like the only way to put a rest above a note is to use voices" - I would reword this, as this makes it sound like there should be some other way. There shouldn't - this is exactly what multiple voices are for - not just in MuseScore, but in any music notation program, and in music notation period.

To be clear: the only way standard music notation has of representing a note and a rest at the same time is using multiple voices. It's not some weird arbitrary limitation of MuseScore, it is precisely how music notation works. You have two independent rhythms on the same staff - one is a half note followed by a quarter note, the other is a quarter rest followed by a half note. So, yes, absolutely that's two separate voices, not just in MuseScore but in any music notation program and in any music theory textbook.

Other than multiple voices, there is really only one other situation in which music notation would ever allow a rest and a note at the same time on the same staff, and that is cross-staff notation - a case where either the note or rest actually "belongs" to another staff. MuseScore also supports directly, as converted in the corresponding section of the Handbook.

In reply to by dave2020X

This is a very astute observation. Somehow pianists learn to read multiple voices (and cross-staff notation) but it's rarely ever explained in a way that allows us to really grasp what is going on conceptually and what's involved in writing it. It's one of these things that is meant to be "obvious" when done right, but learning how to do it right it another matter.

To me this puts it in a similar category as writing rhythms correctly - eg, to expose beat 3 in 4/4 time. Most of us see this done all the time and kind of take it for granted, but then figuring out how to do it yourself in your own writing comes as a bit of a shock at first.

Do you still have an unanswered question? Please log in first to post your question.