A bar is a unit of time before it is a container for a chord, and the count is what a reader is really tracking. So when two chords have to share four beats, the notation does not give you two boxes — it divides one, and the bar count on the page stays honest.
Three of the bars below are split, and they are not split the same way. Two of them cut four beats down the middle, two apiece. One of them gives three beats to the first chord and one to the second, which is a genuinely different feel and is the reason the beats are named rather than assumed.
Below, the same eight bars in D and then in G — a fourth up, the single most common move a session makes to a written page. The middle column is what is actually printed, and it is the column that does not change.
| Bar | Written | In D | In G |
|---|---|---|---|
| 1 | 1 | D | G |
| 2 | 4 5 | G A | C D |
| 3 | 6- | Bm | Em |
| 4 | 4 5 | G A | C D |
| 5 | 1 | D | G |
| 6 | 2-7 57 | Em7 A7 | Am7 D7 |
| 7 | 1 | D | G |
| 8 | 1 | D | G |
The arithmetic that makes it readable
Every bar on this page adds up to four beats, and the build checks all of them. That sounds like a formality until you have played from a chart where it was not true: a bar that quietly contains five beats of chords is not a stylistic choice, it is a page that will stop a band mid-take while four people work out which of them is wrong.
An uneven split — three beats then one — is where a reader most often loses the count, because the eye reads two symbols and the hand assumes two beats. There is no trick for it beyond having met one before, which is what this chart is for.
The sixth bar is the one to look at twice. It splits a minor seventh into a dominant seventh, which is the single most common two-chord bar in the idiom, and the two marks on it are doing different jobs: the minus on the first changes the chord, the seven on the second changes what it is pulling towards. The legend has both.