diff --git a/recipes/atlantic.recipe b/recipes/atlantic.recipe
index 54449b2412..83a006f8e0 100644
--- a/recipes/atlantic.recipe
+++ b/recipes/atlantic.recipe
@@ -5,9 +5,14 @@ __copyright__ = '2008, Kovid Goyal <kovid at kovidgoyal.net>'
 '''
 theatlantic.com
 '''
-import re
+import html5lib
+from lxml import html
 from calibre.web.feeds.news import BasicNewsRecipe
 
+def classes(classes):
+    q = frozenset(classes.split(' '))
+    return dict(attrs={'class':lambda x:x and frozenset(x.split()).intersection(q)})
+
 class TheAtlantic(BasicNewsRecipe):
 
     title      = 'The Atlantic'
@@ -18,7 +23,7 @@ class TheAtlantic(BasicNewsRecipe):
     encoding = 'utf-8'
 
     keep_only_tags = [
-        {'attrs':{'class':['article-header', 'article-body', 'article-magazine', 'metadata', 'article-cover-content']}},
+        classes('article-header article-body article-magazine metadata article-cover-content lead-img'),
     ]
     remove_tags        = [
         {'name': ['meta', 'link', 'noscript']},
@@ -26,12 +31,18 @@ class TheAtlantic(BasicNewsRecipe):
         {'attrs':{'class':lambda x: x and 'article-tools' in x}},
         {'src':lambda x:x and 'spotxchange.com' in x},
     ]
+    remove_tags_after = classes('article-body')
+
     no_stylesheets = True
     remove_attributes = ['style']
-    preprocess_regexps = [
-        (re.compile(r'<script\b.+?</script>', re.DOTALL), lambda m: ''),
-        (re.compile(r'^.*<html', re.DOTALL|re.IGNORECASE), lambda m: '<html'),
-    ]
+
+    def get_browser(self):
+        br = BasicNewsRecipe.get_browser(self)
+        br.set_cookie('inEuropeanUnion', '0', '.theatlantic.com')
+        return br
+
+    def preprocess_raw_html(self, raw, url):
+        return html.tostring(html5lib.parse(raw, treebuilder='lxml', namespaceHTMLElements=False), method='html', encoding=unicode)
 
     def print_version(self, url):
         return url + '?single_page=true'