diff --git a/src/parser/latex.rs b/src/parser/latex.rs
index d1c88469..6b6f82b6 100644
--- a/src/parser/latex.rs
+++ b/src/parser/latex.rs
@@ -18,7 +18,9 @@ use crate::sentence::unicode::{
 // Overleaf leftover names (tokens.mjs) are `IEEEeqnarray` /
 // `IEEEeqnarray*` / `subeqnarray` / `subeqnarray*` / `xltabular` /
 // `math*`. `tikzcd` / pgfplots `axis` / `pgfpicture` are the same
-// class (and starred variants). Not every pgfplots name.
+// class (and starred variants). `groupplot`, `smithchart`
+// (alias of `smithchartaxis`), `polaraxis`, and `ternaryaxis`
+// are that class too. Not every pgfplots name.
 // mathtools `multlined` / `lgathered` / `rgathered` and the starred
 // matrix family, plus breqn `dmath` / `dmath*`, are the same class.
 // There is no `multlined*` / `lgathered*` / `rgathered*`. breqn
@@ -57,6 +59,7 @@ static NON_PROSE_ENVS: &[&str] = &[
     "eqnarray*",
     "IEEEeqnarray",
     "IEEEeqnarray*",
+    "IEEEeqnarraybox",
     "subeqnarray",
     "subeqnarray*",
     "split",
@@ -100,6 +103,7 @@ static NON_PROSE_ENVS: &[&str] = &[
     "verbatim",
     "minted",
     "tikzpicture",
+    "tikzpicture*",
     "tikzcd",
     "tikzcd*",
     "quantikz",
@@ -108,6 +112,14 @@ static NON_PROSE_ENVS: &[&str] = &[
     "pgfpicture*",
     "axis",
     "axis*",
+    "loglogaxis",
+    "semilogxaxis",
+    "semilogyaxis",
+    "groupplot",
+    "smithchart",
+    "smithchartaxis",
+    "polaraxis",
+    "ternaryaxis",
     "array",
     "array*",
     "subarray",
@@ -115,6 +127,10 @@ static NON_PROSE_ENVS: &[&str] = &[
     "prooftree",
     "empheq",
     "empheq*",
+    "dgroup",
+    "dgroup*",
+    "darray",
+    "darray*",
     "matrix",
     "matrix*",
     "pmatrix",
@@ -320,6 +336,12 @@ fn is_builtin_code_env(name: &str) -> bool {
             | "VerbEnv"
             | "alltt"
             | "boxedverbatim"
+            | "semiverbatim"
+            | "verbbox"
+            | "myverbbox"
+            | "verbnobox"
+            | "markdown"
+            | "markdown*"
             | "verbatimtab"
             | "verbatimwrite"
             | "verbwrite"
diff --git a/src/parser/pandoc/shield.rs b/src/parser/pandoc/shield.rs
index 74a702b7..9f075a19 100644
--- a/src/parser/pandoc/shield.rs
+++ b/src/parser/pandoc/shield.rs
@@ -234,8 +234,15 @@ fn toggle_code(
             || lower.starts_with("#+begin_example")
             || lower.starts_with("#+begin_export")
             || lower.starts_with("\\begin{verbatim}")
+            || lower.starts_with("\\begin{verbatim*}")
+            || lower.starts_with("\\begin{bverbatim}")
             || lower.starts_with("\\begin{lstlisting}")
+            || lower.starts_with("\\begin{lstlisting*}")
             || lower.starts_with("\\begin{minted}")
+            || lower.starts_with("\\begin{minted*}")
+            || lower.starts_with("\\begin{alltt}")
+            || lower.starts_with("\\begin{luacode}")
+            || lower.starts_with("\\begin{luacode*}")
         {
             *in_code = true;
             *fence_mark = Some(lower.chars().take(24).collect());
@@ -256,8 +263,15 @@ fn toggle_code(
         || lower.starts_with("#+end_example")
         || lower.starts_with("#+end_export")
         || lower.starts_with("\\end{verbatim}")
+        || lower.starts_with("\\end{verbatim*}")
+        || lower.starts_with("\\end{bverbatim}")
         || lower.starts_with("\\end{lstlisting}")
+        || lower.starts_with("\\end{lstlisting*}")
         || lower.starts_with("\\end{minted}")
+        || lower.starts_with("\\end{minted*}")
+        || lower.starts_with("\\end{alltt}")
+        || lower.starts_with("\\end{luacode}")
+        || lower.starts_with("\\end{luacode*}")
     {
         *in_code = false;
         *fence_mark = None;
@@ -316,6 +330,30 @@ mod tests {
         assert!(!chunk.contains("After."));
     }
 
+    #[test]
+    fn pragma_inside_starred_verbatim_is_ignored() {
+        for env in ["verbatim*", "lstlisting*", "minted*"] {
+            let input =
+                format!("\\begin{{{env}}}\nsnapper:off\nSee Dr. Smith. He left.\n\\end{{{env}}}\n");
+            assert!(
+                pragma_spans(&input).is_empty(),
+                "{env} must hide an interior pragma, got spans in {input:?}"
+            );
+        }
+    }
+
+    #[test]
+    fn pragma_inside_alltt_bverbatim_and_luacode_is_ignored() {
+        for env in ["alltt", "BVerbatim", "luacode", "luacode*"] {
+            let input =
+                format!("\\begin{{{env}}}\nsnapper:off\nSee Dr. Smith. He left.\n\\end{{{env}}}\n");
+            assert!(
+                pragma_spans(&input).is_empty(),
+                "{env} must hide an interior pragma, got spans in {input:?}"
+            );
+        }
+    }
+
     #[test]
     fn pragma_inside_markdown_fence_is_ignored() {
         let input = "```\n<!-- snapper:off -->\ncode\n<!-- snapper:on -->\n```\n";
diff --git a/src/sentence/unicode.rs b/src/sentence/unicode.rs
index ad6836d3..33c54e41 100644
--- a/src/sentence/unicode.rs
+++ b/src/sentence/unicode.rs
@@ -3513,6 +3513,22 @@ fn wrap_closes_after_bang(piece: &str, bang_at: usize) -> bool {
     !dq && !sq && latex <= 0 && paren <= 0 && bracket <= 0 && brace <= 0
 }
 
+/// Append `` `Capital `` onto a segment that already ends in `.!?`.
+/// No space: a space is a new UAX boundary.
+fn glue_tick_capital_after_punct(result: &mut Vec<String>, segment: &str) -> bool {
+    if !result.last().is_some_and(|last| {
+        last.ends_with(['.', '!', '?']) && piece_starts_sentence_after_ticks(segment)
+    }) {
+        return false;
+    }
+    if let Some(last) = result.last_mut() {
+        last.push_str(segment);
+    } else {
+        result.push(segment.to_string());
+    }
+    true
+}
+
 fn merge_splits_inside_delimiters(segments: Vec<String>) -> Vec<String> {
     let mut result: Vec<String> = Vec::with_capacity(segments.len());
     let mut state = DelimState::default();
@@ -3532,10 +3548,10 @@ fn merge_splits_inside_delimiters(segments: Vec<String>) -> Vec<String> {
                 }
                 continue;
             }
-            if result.last().is_some_and(|last| {
-                last.ends_with(['.', '!', '?']) && piece_starts_sentence_after_ticks(&segment)
-            }) {
-                result.push(segment.clone());
+            // `.` + leftover `` `Capital `` is a UAX break. A newline here
+            // sits inside the open span (`""`A`"`a`.`A"`). Glue with no
+            // invented space.
+            if glue_tick_capital_after_punct(&mut result, &segment) {
                 state.feed(&segment);
                 continue;
             }
@@ -3544,6 +3560,11 @@ fn merge_splits_inside_delimiters(segments: Vec<String>) -> Vec<String> {
             } else {
                 result.push(segment.clone());
             }
+        } else if glue_tick_capital_after_punct(&mut result, &segment) {
+            // `=(a=` is an org span once a newline precedes `=`, so the
+            // paren is hidden and this arm must glue the same ticks.
+            state.feed(&segment);
+            continue;
         } else {
             result.push(segment.clone());
         }
diff --git a/tests/latex_more_envs.rs b/tests/latex_more_envs.rs
new file mode 100644
index 00000000..16922f93
--- /dev/null
+++ b/tests/latex_more_envs.rs
@@ -0,0 +1,102 @@
+//! Math and verbatim environments whose bodies must not reflow as prose.
+//! Surrounding prose still splits.
+
+use snapper_fmt::format::Format;
+use snapper_fmt::parser::latex::LatexParser;
+use snapper_fmt::parser::{FormatParser, Region};
+use snapper_fmt::{FormatConfig, format_text};
+
+fn latex_cfg() -> FormatConfig {
+    FormatConfig {
+        format: Format::Latex,
+        ..Default::default()
+    }
+    .without_safety_backstops()
+}
+
+fn structure_env_does_not_reflow(name: &str, begin_suffix: &str) {
+    let input = format!(
+        "\\begin{{{name}}}{begin_suffix}\nThis is a long sentence that must stay inside {name} and must not reflow as prose. Another sentence stays put.\n\\end{{{name}}}\nAfter the env. Second sentence.\n"
+    );
+    let fused = format!(
+        "This is a long sentence that must stay inside {name} and must not reflow as prose. Another sentence stays put."
+    );
+    let regions = LatexParser::default().parse(&input);
+    assert!(
+        regions
+            .iter()
+            .any(|r| matches!(r, Region::Structure(s) if s.contains(&fused))),
+        "{name} body must be Structure, got: {regions:?}"
+    );
+    assert!(
+        !regions
+            .iter()
+            .any(|r| matches!(r, Region::Prose(p) if p.contains("must not reflow"))),
+        "{name} body must not be Prose, got: {regions:?}"
+    );
+    let out = format_text(&input, &latex_cfg()).unwrap();
+    assert!(
+        out.contains(&fused),
+        "{name} body must stay one source line, got:\n{out}"
+    );
+    assert!(
+        out.contains("After the env.\nSecond sentence."),
+        "prose after {name} must still reflow, got:\n{out}"
+    );
+    assert_eq!(format_text(&out, &latex_cfg()).unwrap(), out);
+}
+
+fn code_env_does_not_reflow(name: &str, begin_suffix: &str) {
+    let input =
+        format!("\\begin{{{name}}}{begin_suffix}\nFirst. Second.\n\\end{{{name}}}\nAfter. Next.\n");
+    let regions = LatexParser::default().parse(&input);
+    assert!(
+        regions.iter().any(|r| matches!(
+            r,
+            Region::Code { body, .. } if body.contains("First. Second.")
+        )),
+        "{name} body must be Code, got: {regions:?}"
+    );
+    let out = format_text(&input, &latex_cfg()).unwrap();
+    assert!(
+        out.contains("First. Second."),
+        "{name} body must stay one source line, got:\n{out}"
+    );
+    assert!(
+        !out.contains("First.\nSecond."),
+        "{name} must not reflow as prose, got:\n{out}"
+    );
+    assert!(
+        out.contains("After.\nNext."),
+        "prose after {name} must still split, got:\n{out}"
+    );
+    assert_eq!(format_text(&out, &latex_cfg()).unwrap(), out);
+}
+
+#[test]
+fn more_math_envs_stay_structure() {
+    structure_env_does_not_reflow("IEEEeqnarraybox", "{rCl}");
+    structure_env_does_not_reflow("tikzpicture*", "");
+    structure_env_does_not_reflow("loglogaxis", "");
+    structure_env_does_not_reflow("semilogxaxis", "");
+    structure_env_does_not_reflow("semilogyaxis", "");
+    structure_env_does_not_reflow("groupplot", "");
+    structure_env_does_not_reflow("smithchart", "");
+    structure_env_does_not_reflow("smithchartaxis", "");
+    structure_env_does_not_reflow("polaraxis", "");
+    structure_env_does_not_reflow("ternaryaxis", "");
+    structure_env_does_not_reflow("dgroup", "");
+    structure_env_does_not_reflow("dgroup*", "");
+    structure_env_does_not_reflow("darray", "");
+    structure_env_does_not_reflow("darray*", "");
+}
+
+#[test]
+fn more_verbatim_envs_stay_code() {
+    code_env_does_not_reflow("semiverbatim", "");
+    code_env_does_not_reflow("verbbox", "");
+    code_env_does_not_reflow("myverbbox", "{\\mybox}");
+    code_env_does_not_reflow("verbnobox", "");
+    code_env_does_not_reflow("markdown", "");
+    code_env_does_not_reflow("markdown*", "");
+}
diff --git a/tests/sentence_delim_props.rs b/tests/sentence_delim_props.rs
index e93b0ad4..059f0c9d 100644
--- a/tests/sentence_delim_props.rs
+++ b/tests/sentence_delim_props.rs
@@ -150,6 +150,31 @@ fn plaintext_period_before_backticks_quote_is_span_safe() {
     );
 }
 
+/// macOS proptest seed. `""`A`"`a`.`A"` is delimiters_balanced.
+/// The period before `` `A" `` must not become a line break inside the quote.
+#[test]
+fn plaintext_balanced_quote_period_backtick_is_span_safe() {
+    let input = "\"\"`A`\"`a`.`A\"";
+    assert!(
+        delimiters_balanced(&input),
+        "seed must be balanced so the property would fire\n in={input:?}"
+    );
+    let out = format_plain(&input);
+    let again = format_plain(&out);
+    assert_eq!(
+        out, "\"\"`A`\"`a`.`A\"\n",
+        "must not split after the period\n in={input:?}\n out={out:?}"
+    );
+    assert_eq!(
+        again, out,
+        "idempotence\n in={input:?}\n out={out:?}\n again={again:?}"
+    );
+    assert!(
+        newlines_respect_delimiter_spans(&out),
+        "span newline\n in={input:?}\n out={out:?}"
+    );
+}
+
 /// GitHub #374. `".` A`'``aaa0`"0` is delimiters_balanced; format_text
 /// used to insert a newline after `.` so newlines_respect_delimiter_spans
 /// failed. Markup closer-split must not break the still-open quote.