diff --git a/exercises/crypto-square/README.md b/exercises/crypto-square/README.md
index c1c6dba93..07830ff53 100644
--- a/exercises/crypto-square/README.md
+++ b/exercises/crypto-square/README.md
@@ -47,9 +47,10 @@ imtgdvsfearwermayoogoanouuiontnnlvtwttddesaohghnsseoau
 
 Output the encoded text in chunks.  Phrases that fill perfect rectangles
 `(r X c)` should be output `c` chunks of `r` length, separated by spaces.
-Phrases that do not fill perfect rectangles will have `n` empty spaces.
-Those spaces should be distributed evenly, added to the end of the last
-`n` chunks.
+Phrases that do not fill perfect rectangles will be `n` characters short.
+To minimize the jaggedness, the missing chars should be distributed evenly,
+added to the end of the last `n` chunks. The bottom row(s) should be, at most,
+one char short of the top row(s).
 
 ```text
 imtgdvs fearwer mayoogo anouuio ntnnlvt wttddes aohghn  sseoau 
diff --git a/exercises/crypto-square/tests/crypto-square.rs b/exercises/crypto-square/tests/crypto-square.rs
index 087de1b60..20d41851d 100644
--- a/exercises/crypto-square/tests/crypto-square.rs
+++ b/exercises/crypto-square/tests/crypto-square.rs
@@ -63,6 +63,14 @@ fn test_ignore_non_ascii_chars() {
     );
 }
 
+#[test]
+#[ignore]
+fn test_minimize_jaggedness() {
+    test("aaaa bbbb cccc d", "abcd abc abc abc");
+    test("aaaa bbbb cccc dd", "abcd abcd abc abc");
+    test("aaaa bbbb cccc ddd", "abcd abcd abcd abc");
+}
+
 #[test]
 #[ignore]
 fn test_long() {