diff --git a/crates/ui-egui/src/i18n/cs.tsv b/crates/ui-egui/src/i18n/cs.tsv
index 67d5709b77..3d287457e5 100644
--- a/crates/ui-egui/src/i18n/cs.tsv
+++ b/crates/ui-egui/src/i18n/cs.tsv
@@ -5657,7 +5657,7 @@ mark-style	Japanese	Japonské
 	Word Spacing:	Mezery mezi slovy:
 	Letter Spacing:	Mezery mezi písmeny:
 	Glyph Scaling:	Změna velikosti glyfů:
-	Auto Leading:	Automatické proklady:
+	Auto Leading:	Automatický proklad:
 	Single Word Justification:	Zarovnání jednoho slova:
 	Full Justify	Zarovnat do bloku
 	Minimum must not exceed Desired, nor Desired Maximum.	Minimum nesmí překročit požadovanou hodnotu ani požadovaná hodnota maximum.
diff --git a/crates/ui-egui/src/i18n/tests.rs b/crates/ui-egui/src/i18n/tests.rs
index bb2de38d61..2fe5cb4aac 100644
--- a/crates/ui-egui/src/i18n/tests.rs
+++ b/crates/ui-egui/src/i18n/tests.rs
@@ -524,9 +524,36 @@ fn zh_hans_is_simplified_and_mainland() {
     assert!(bad.is_empty(), "in zh-hans.tsv:\n{}", bad.join("\n"));
 }
 
+/// Whether `text` holds `term`. In a script that separates words with spaces the term must stand as
+/// whole words, so "liderlik" does not hold "lider"; in Chinese, Japanese, Thai, Lao, Khmer and
+/// Myanmar text any occurrence counts.
+fn contains_term(text: &str, term: &str) -> bool {
+    let unspaced = |c: char| matches!(c as u32, 0x0E00..=0x0EFF | 0x1000..=0x109F | 0x1780..=0x17FF | 0x3040..=0x30FF | 0x3400..=0x4DBF | 0x4E00..=0x9FFF | 0xF900..=0xFAFF);
+    if term.chars().any(unspaced) {
+        return text.contains(term);
+    }
+    text.match_indices(term).any(|(at, m)| {
+        !text[..at].chars().next_back().is_some_and(char::is_alphanumeric) && !text[at + m.len()..].chars().next().is_some_and(char::is_alphanumeric)
+    })
+}
+
+#[test]
+fn contains_term_matches_whole_words_where_words_are_spaced() {
+    assert!(contains_term("tepeden tepeye lider", "lider"));
+    assert!(contains_term("boyut/lider:", "lider"));
+    assert!(contains_term("réinitialiser l’approche", "approche"));
+    assert!(!contains_term("tepeden tepeye liderlik", "lider"));
+    assert!(!contains_term("automatické proklady:", "proklad"));
+    assert!(contains_term("kích thước: giãn chữ:", "giãn chữ"));
+    assert!(!contains_term("kích thước chữ: giãn:", "giãn chữ"));
+    assert!(contains_term("自動行送り:", "行送り"));
+    assert!(contains_term("大小：字距微調：", "字距"));
+}
+
 /// Each catalog names leading with one term: every word of the Character panel's Leading field
-/// (the text before " (", ignoring case) appears in the Auto Leading, Size/Leading, Top-to-Top
-/// Leading and Bottom-to-Bottom Leading rows outside their placeholders.
+/// (the text before " (", ignoring case) appears as a whole word (see [`contains_term`]) in the
+/// Auto Leading, Size/Leading, Top-to-Top Leading and Bottom-to-Bottom Leading rows outside their
+/// placeholders.
 /// A catalog without the Leading row, or without one of the others, is skipped for that check.
 #[test]
 fn leading_term_agrees_within_each_catalog() {
@@ -540,7 +567,7 @@ fn leading_term_agrees_within_each_catalog() {
         for src in ["Auto Leading:", "Size/Leading", "Top-to-Top Leading", "Bottom-to-Bottom Leading"] {
             if let Some(text) = row(src) {
                 let bare = placeholders(&text).iter().fold(text.clone(), |t, p| t.replace(&format!("{{{p}}}"), " "));
-                for word in leading.split_whitespace().filter(|w| !bare.contains(*w)) {
+                for word in leading.split_whitespace().filter(|w| !contains_term(&bare, w)) {
                     bad.push(format!("{}: {src:?} → {text:?} lacks {word:?} from the Leading field {leading:?}", l.code));
                 }
             }
@@ -576,10 +603,12 @@ fn no_translated_word_mixes_latin_cyrillic_and_greek_letters() {
 }
 
 /// Each catalog gives the Character panel's Kerning, Leading and Tracking fields three different
-/// names, every word of the Tracking and Kerning names appears in the Reset Tracking and Kerning
-/// row, and every word of the Tracking name appears in the Info panel's type line outside its
-/// placeholders (ignoring case). Each word is matched once, the longer name first, so a name
-/// that contains the other name does not count for both.
+/// names, and every word of the Tracking and Kerning names appears in the Reset Tracking and
+/// Kerning row (ignoring case). Each word is matched once, the longer name first, so a name that
+/// contains the other name does not count for both. The Info panel's type line labels
+/// `{tracking}` with the whole Tracking name: the text between the placeholder before it and
+/// `{tracking}` holds the name (see [`contains_term`]) once any Kerning or Leading name that
+/// contains it has been taken out.
 /// A catalog without one of these rows is skipped for the checks that need it.
 #[test]
 fn character_panel_terms_agree_within_each_catalog() {
@@ -613,7 +642,16 @@ fn character_panel_terms_agree_within_each_catalog() {
             }
         };
         uses("Reset Tracking and Kerning", &[&tracking, &kerning]);
-        uses("Size: {size}  Tracking: {tracking}", &[&tracking]);
+        if let (Some(text), Some(tracking)) = (row("Size: {size}  Tracking: {tracking}"), &tracking) {
+            let before = text.split("{tracking}").next().unwrap_or_default();
+            let mut label = before.rsplit('}').next().unwrap_or_default().to_string();
+            for other in [&kerning, &leading].into_iter().flatten().filter(|o| *o != tracking && o.contains(tracking.as_str())) {
+                label = label.replace(other.as_str(), " ");
+            }
+            if !contains_term(&label, tracking) {
+                bad.push(format!("{}: the Info row {text:?} labels {{tracking}} {label:?}, not the field {tracking:?}", l.code));
+            }
+        }
     }
     assert!(bad.is_empty(), "Character panel terms that disagree:\n{}", bad.join("\n"));
 }