Partially solved Nonograms often contain two filled fragments with unknown cells between them. The key question is whether those fragments must become one clue block or must remain separate.
That decision is never based on how the picture looks. It follows from clue lengths and legal block assignments.
Joining: one clue must cover both fragments
Consider a 9-cell line with a single clue 5:
??■■?■■??
There is only one block in the entire line, so both filled fragments must belong to that same block.
A length-5 block covering both fragments must span cells 3–7. The unknown cell between them is therefore forced filled.
The two fragments join into one continuous block of five.
Splitting: joining would violate every compatible clue size
Now consider a 10-cell line with clues 3 3 and known filled fragments at cells 2–3 and 5–6:
?■■?■■????
If cell 4 were filled, cells 2–6 would form a continuous run of five filled cells. But the only clue blocks have length 3.
A five-cell block cannot exist, so cell 4 must be empty. The two fragments must remain in separate clue blocks.
Once the separator is known, each fragment can be extended only within its own side of that boundary.
What proves that fragments must join?
Fragments can be forced to join when:
- only one unresolved clue block can cover both;
- a segment is assigned to a single clue and both fragments lie inside its legal reach;
- separating the fragments would require more clue blocks than remain;
- clue order makes every separate assignment impossible.
The proof is always that every valid line pattern connects them.
What proves that fragments must split?
Fragments must stay separate when:
- joining them would create a run longer than every compatible clue;
- a confirmed X already lies between them;
- clue assignments prove they belong to different clue blocks;
- joining them would leave no legal place for another required block;
- clue order would be violated.
A split often gives you a new separator immediately.
Do not confuse “close together” with “same block”
Two filled cells separated by one unknown cell are visually tempting to join. But that unknown may need to be empty.
Similarly, fragments several cells apart may still belong to one long clue if all cells between them are eventually forced filled.
Distance alone proves nothing.
Use assigned clue identity whenever possible
Joining and splitting becomes much easier after clue-to-segment assignment.
If you know a segment contains only clue 7, every filled fragment in that segment must be covered by the same seven-cell block.
If you know two fragments belong to consecutive clues 3 and 4, there must be at least one empty cell separating them.
Sometimes neither joining nor splitting is proven yet
A useful advanced skill is recognizing when the evidence is insufficient.
Consider a 12-cell line with clues 2 5 and confirmed fills at cells 5 and 7:
????■?■?????
There are legal patterns in which both filled cells belong to the later 5-block. There are also legal patterns in which cell 5 belongs to the first 2-block and cell 7 belongs to the later 5-block, with a separator between them.
So neither conclusion is justified yet:
- filling cell 6 to join the fragments would be premature;
- marking cell 6 empty to split them would also be premature.
The correct state of cell 6 is still unknown until another constraint removes one family of patterns.
This example matters because advanced mistakes often come from turning a plausible fragment identity into a fact.
Fragment identity can change before the clues change
The clue sequence is fixed, but your interpretation of a particular filled fragment may remain flexible.
A fragment can sometimes be compatible with clue 2 in one valid pattern and clue 5 in another. When a crossing-line X later removes one of those patterns, the fragment's clue identity suddenly becomes fixed, and several joining/splitting deductions can follow at once.
That creates a useful workflow:
- record which clues can still own each fragment;
- use segment assignment and placement bounds to remove impossible owners;
- only then decide whether neighboring fragments must connect or separate.
Joining and splitting are statements about every valid pattern
The safest formal definitions are:
- fragments must join if every valid line pattern assigns them to one continuous clue block;
- fragments must split if every valid line pattern assigns them to different clue blocks;
- the relationship is unresolved if both kinds of pattern still exist.
This framing prevents local visual intuition from outrunning the actual constraints.
Step-by-step fragment test
When two filled fragments might interact:
- identify which unresolved clues could cover each fragment;
- test whether one legal clue block can cover both without exceeding its length;
- test whether they can belong to different blocks while preserving clue order and separators;
- reject any case that conflicts with known X marks or segment boundaries;
- fill the cells between fragments only if every valid case joins them;
- mark a separating cell empty only if every valid case keeps them apart.
Common mistakes
Joining fragments because the hidden image “needs a line”
The picture is not a clue.
Forgetting the clue's maximum run length
A continuous filled run cannot become longer than the clue block assigned to it.
Splitting fragments without room for a mandatory separator
Different monochrome blocks require at least one empty cell between them.
Assuming fragment identity before clue assignment is known
On advanced lines, the same fragment may still be compatible with more than one clue. Keep alternatives until one is ruled out.
What to learn next
Once block identities become clearer, use block order and placement bounds to calculate the earliest and latest legal positions for each clue block rather than treating the line as one undifferentiated set of possibilities.
FAQ
Can two filled fragments belong to the same clue block?
Yes, if all cells between them can be filled and the resulting continuous run matches that clue's legal length and position.
Can two adjacent filled fragments belong to different monochrome clues?
No. Different monochrome clue blocks require at least one empty separator, so adjacent filled cells form one continuous block.
Is joining the same as extending a confirmed block?
They are related. Extension grows one known fragment; joining proves that two fragments are parts of the same block and fills the bridge between them.