diff --git a/src/exercises/day-1/for-loops.md b/src/exercises/day-1/for-loops.md
index 86a5ef18..c4824cb7 100644
--- a/src/exercises/day-1/for-loops.md
+++ b/src/exercises/day-1/for-loops.md
@@ -73,3 +73,10 @@ slice-of-slices. Why or why not?
 
 See the [`ndarray` crate](https://docs.rs/ndarray/) for a production quality
 implementation.
+
+<details>
+
+The solution and the answer to the bonus section are available in the 
+[Solution](solutions-morning.md#arrays-and-for-loops) section.
+
+</details>
diff --git a/src/exercises/day-1/solutions-morning.md b/src/exercises/day-1/solutions-morning.md
index cd13ad56..d5aed0db 100644
--- a/src/exercises/day-1/solutions-morning.md
+++ b/src/exercises/day-1/solutions-morning.md
@@ -11,4 +11,6 @@
 
 It honestly doesn't work so well. It might seem that we could use a slice-of-slices (`&[&[i32]]`) as the input type to transpose and thus make our function handle any size of matrix. However, this quickly breaks down: the return type cannot be `&[&[i32]]` since it needs to own the data you return.
 
-You can attempt to use something like `Vec<Vec<i32>>`, but this doesn't work very well either: it's hard to convert from `Vec<Vec<i32>>` to `&[&[i32]]` so now you cannot easily use `pretty_print` either.
\ No newline at end of file
+You can attempt to use something like `Vec<Vec<i32>>`, but this doesn't work very well either: it's hard to convert from `Vec<Vec<i32>>` to `&[&[i32]]` so now you cannot easily use `pretty_print` either.
+
+In addition, the type itself would not enforce that the child slices are of the same length, so such variable could contain an invalid matrix.