diff --git a/CHANGELOG.md b/CHANGELOG.md
index 63f0e21..13b2530 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -29,6 +29,10 @@ Versionsnummern nach [Semantic Versioning](https://semver.org/lang/de/).
 
 ### Behoben
 
+- CSS-Auswertung: Es zählen nur noch Stylesheets, die ein Dokument per `<link>` (oder `@import`) einbindet.
+  Bisher wirkten alle CSS-Dateien des Manifests, auch mitgelieferte Layouts anderer Buchreihen. Im C-Buch
+  machte das aus fetten Buchstaben Inline-Code (`` `st`an`d`ard ``) und setzte falsche `<mark>`; eine
+  fremde `display:none`-Regel hätte Text gelöscht (QA3-01, Regressionstests in `tests/test_styles.py`).
 - `pandoc/callouts.lua`: Die Kästen blieben ohne Rahmen, wenn man Pandoc zusätzlich `-H`/`--include-in-header`
   übergab, und im EPUB hatten nur die Titelseite, nicht die Kapitel, das Kasten-CSS. Das CSS steht jetzt in
   `pandoc/callouts.css`; der Filter setzt es bei HTML an den Anfang des Dokuments und fügt es bei EPUB den
diff --git a/README.md b/README.md
index 3e07c04..b5ea70c 100644
--- a/README.md
+++ b/README.md
@@ -249,6 +249,8 @@ weil CommonMark keine `[^…]`-Syntax kennt.
 - **CSS-Formatierung:** Viele Verlage setzen Hervorhebungen nur über Stylesheets. epub2md wertet die
   CSS-Dateien des Buches aus: Klassen-, Element- und ID-Selektoren, Nachfahren- und Kind-Kombinatoren,
   Spezifität und `style`-Attribute. `@media`-Blöcke und Selektoren mit Pseudoklassen werden ignoriert.
+  Berücksichtigt werden nur Stylesheets, die ein Dokument per `<link>` (oder `@import`) einbindet;
+  mitgelieferte Layouts anderer Buchreihen bleiben wie im E-Book-Reader wirkungslos (Info `css-unlinked`).
   Direkt aneinanderstoßende gleiche Hervorhebungen werden zusammengefasst (`*x**y*` → `*xy*`).
 - **Farben und Schriften:** Markdown kennt keine Farben, Schriftarten oder -größen. Mit `--styles html`
   bleiben sie als HTML-Stil erhalten. Übernommen wird nur, was vom Grundstil abweicht (häufigste Werte aller
diff --git a/epub2md/pipeline.py b/epub2md/pipeline.py
index bc10e24..9f1f68e 100644
--- a/epub2md/pipeline.py
+++ b/epub2md/pipeline.py
@@ -13,6 +13,7 @@ from dataclasses import asdict, dataclass, field
 from datetime import datetime, timezone
 from pathlib import Path
 from typing import Any
+from urllib.parse import unquote
 
 from . import TOOL_NAME, __version__
 from .chapters import ChapterBuilder
@@ -27,7 +28,7 @@ from .media import MediaRegistry
 from .models import Chapter
 from .normalize import Doc, book_uses_title_tables, is_html_toc_doc, make_doc, normalize_doc, remove_keep_tail
 from .references import LinkResolver
-from .utils import is_element, is_inside, iter_elements, local, sha256_file, slugify, text_content
+from .utils import is_element, is_inside, iter_elements, local, resolve_href, sha256_file, slugify, text_content
 from . import writer
 
 MARKER_FILES = ("conversion-report.json", "manifest.json")
@@ -233,18 +234,39 @@ def _build_structure(docs: list[Doc], package: Any, opts: Options,
     return notes, chapters, resolver
 
 
+CSS_IMPORT = re.compile(r"""@import\s+(?:url\(\s*)?["']?([^"')\s;]+)""", re.I)
+
+
 def _load_stylesheet(reader: EpubReader, package: Any, docs: list[Doc], diag: Diagnostics) -> StyleSheet:
-    """All CSS of the book: manifest stylesheets (manifest order) plus <style> elements of the documents."""
-    texts: list[str] = []
-    for item in package.manifest.values():
-        if item.media_type == "text/css" or item.path.lower().endswith(".css"):
-            try:
-                texts.append(decode_markup(reader.read(item.path), item.path, diag))
-            except (KeyError, ContainerError):
-                diag.warn("css-unreadable", f"Stylesheet {item.path} nicht lesbar; seine Formatierung entfällt.", item.path)
+    """The book's CSS: manifest stylesheets that a spine document links (directly or via @import), in manifest
+    order, plus <style> elements of the documents. Unlinked stylesheets (publishers ship the layouts of other
+    series) are ignored – a reading system would not apply them either."""
+    pending: list[str] = []
     for doc in docs:
-        texts.extend(text_content(e) for e in iter_elements(doc.root) if local(e.tag) == "style")
-    return StyleSheet(texts)
+        for e in iter_elements(doc.root):
+            if local(e.tag) == "link" and "stylesheet" in (e.get("rel") or "").lower().split():
+                pending.append(resolve_href(doc.path, unquote((e.get("href") or "").split("#")[0].split("?")[0])) or "")
+    texts: dict[str, str] = {}
+    while pending:
+        path = pending.pop()
+        if not path or path in texts:
+            continue
+        try:
+            texts[path] = decode_markup(reader.read(path), path, diag)
+        except (KeyError, ContainerError):
+            diag.warn("css-unreadable", f"Stylesheet {path} nicht lesbar; seine Formatierung entfällt.", path)
+            texts[path] = ""
+            continue
+        pending.extend(resolve_href(path, unquote(m)) or "" for m in CSS_IMPORT.findall(texts[path]))
+    css_items = [i.path for i in package.manifest.values() if i.media_type == "text/css" or i.path.lower().endswith(".css")]
+    ordered = [texts[p] for p in css_items if p in texts] + [t for p, t in texts.items() if p not in css_items]
+    unlinked = len([p for p in css_items if p not in texts])
+    if unlinked:
+        diag.info("css-unlinked", f"{unlinked} Stylesheet(s) im Manifest werden von keinem Dokument eingebunden "
+                                  "und daher nicht ausgewertet.")
+    for doc in docs:
+        ordered.extend(text_content(e) for e in iter_elements(doc.root) if local(e.tag) == "style")
+    return StyleSheet(ordered)
 
 
 def _render(reader: EpubReader, package: Any, chapters: list[Chapter], resolver: LinkResolver, notes: Any,
diff --git a/tests/conftest.py b/tests/conftest.py
index a08e5b7..5012e47 100644
--- a/tests/conftest.py
+++ b/tests/conftest.py
@@ -3,6 +3,7 @@
 from __future__ import annotations
 
 import json
+import posixpath
 import sys
 import zipfile
 from dataclasses import dataclass, field
@@ -25,7 +26,7 @@ SVG = b'<svg xmlns="http://www.w3.org/2000/svg" width="10" height="10"><rect wid
 XHTML = """<?xml version="1.0" encoding="utf-8"?>
 <!DOCTYPE html>
 <html xmlns="http://www.w3.org/1999/xhtml" xmlns:epub="http://www.idpf.org/2007/ops" lang="de">
-<head><title>{title}</title></head>
+<head><title>{title}</title>{head}</head>
 <body>{body}</body>
 </html>"""
 
@@ -36,6 +37,7 @@ class Doc:
     body: str
     title: str = ""
     linear: bool = True
+    css: list[str] | None = None  # linked stylesheets (OEBPS-relative) – default: every .css in Book.media
 
 
 @dataclass
@@ -90,7 +92,7 @@ def build_epub(path: Path, book: Book) -> Path:
   <spine{' toc="ncx"' if book.ncx else ''}>{''.join(spine)}{book.spine_extra}</spine>
 </package>"""
     nav_xhtml = XHTML.format(
-        title="Inhalt",
+        title="Inhalt", head="",
         body='<nav epub:type="toc"><ol>' + "".join(f'<li><a href="{h}">{lbl}</a></li>' for lbl, h in nav_entries) + "</ol></nav>",
     )
     ncx = (
@@ -114,7 +116,10 @@ def build_epub(path: Path, book: Book) -> Path:
         if book.ncx:
             z.writestr("OEBPS/toc.ncx", ncx)
         for d in book.docs:
-            z.writestr(f"OEBPS/{d.name}", XHTML.format(title=d.title, body=d.body))
+            css = d.css if d.css is not None else [n for n in book.media if n.lower().endswith(".css")]
+            head = "".join(f'<link rel="stylesheet" type="text/css" href="{posixpath.relpath(c, posixpath.dirname(d.name) or ".")}"/>'
+                           for c in css)
+            z.writestr(f"OEBPS/{d.name}", XHTML.format(title=d.title, head=head, body=d.body))
         for name, data in book.media.items():
             z.writestr(f"OEBPS/{name}", data)
         for name, data in book.extra_entries.items():
diff --git a/tests/test_styles.py b/tests/test_styles.py
index b3faa65..a62a99d 100644
--- a/tests/test_styles.py
+++ b/tests/test_styles.py
@@ -211,3 +211,22 @@ def test_page_marker_in_box_title_goes_before_the_box(convert):
     text = convert(Book(docs=[Doc("c.xhtml", body)]), "--page-markers", "visible").chapter("001")
     assert "\\[S. 62\\]\n\n> [!NOTE]\n> **Wozu Deklarationen?**" in text
     assert "> Weil es \\[S. 63\\]hilft." in text  # markers in the box body stay in place
+
+
+def test_qa3_01_unlinked_stylesheets_are_ignored(convert):
+    """Publishers ship layout CSS of other series in the manifest; only linked stylesheets apply (QA3-01)."""
+    css = {"main.css": b".bold { font-weight: bold; }",
+           "fremd.css": b"p > span.bold { font-weight: normal; font-family: monospace; } .weg { display: none; }"}
+    body = ('<h1>K</h1><p>Die (<span class="bold">st</span>an<span class="bold">d</span>ard '
+            '<span class="bold">in</span>put) und <span class="weg">sichtbarer Text</span>.</p>')
+    text = convert(Book(docs=[Doc("c.xhtml", body, css=["main.css"])], media=css)).chapter("001")
+    assert "`" not in text  # monospace rule of fremd.css must not turn bold into code
+    assert "**st**an**d**ard **in**put" in text
+    assert "sichtbarer Text" in text  # display:none of fremd.css must not hide content
+
+
+def test_qa3_01_imported_stylesheets_still_apply(convert):
+    css = {"main.css": b'@import url("extra.css");', "extra.css": b".k { font-style: italic; }"}
+    body = '<h1>K</h1><p><span class="k">kursiv</span> normal.</p>'
+    text = convert(Book(docs=[Doc("c.xhtml", body, css=["main.css"])], media=css)).chapter("001")
+    assert "*kursiv* normal." in text