diff --git a/lib/coderay/encoders/html.rb b/lib/coderay/encoders/html.rb
index 1b33e921..48b8696d 100644
--- a/lib/coderay/encoders/html.rb
+++ b/lib/coderay/encoders/html.rb
@@ -249,7 +249,7 @@ def begin_line kind
         if style['class="']
           @out << style.sub('class="', 'class="line ')
         else
-          @out << style.sub('>', ' class="line">')
+          @out << style.sub(/>\s*\z/, ' class="line">')
         end
       else
         @out << '<span class="line">'
diff --git a/test/unit/html_escaping.rb b/test/unit/html_escaping.rb
new file mode 100644
index 00000000..7b4c8640
--- /dev/null
+++ b/test/unit/html_escaping.rb
@@ -0,0 +1,80 @@
+require 'test/unit'
+require 'coderay'
+
+class HtmlEscapingTest < Test::Unit::TestCase
+  
+  def test_begin_line_with_greater_than_in_attributes
+    # Test that begin_line correctly handles '>' characters in HTML attributes
+    # This addresses issue #263: Incomplete string escaping or encoding
+    
+    encoder = CodeRay::Encoders::HTML.new
+    encoder.instance_variable_set(:@out, '')
+    encoder.instance_variable_set(:@opened, [])
+    encoder.instance_variable_set(:@span_for_kinds, {
+      # Simulate a span with '>' characters in title attribute
+      :test => '<span title="value > other" class="keyword">'
+    })
+    
+    # Call begin_line which should add class="line" without breaking the title attribute
+    encoder.send(:begin_line, :test)
+    
+    result = encoder.instance_variable_get(:@out)
+    
+    # The result should have class="line" added to the existing class attribute
+    # and should NOT have corrupted the title attribute
+    assert_equal '<span title="value > other" class="line keyword">', result
+    assert result.include?('title="value > other"'), "Title attribute should remain intact"
+    assert result.include?('class="line keyword"'), "Class should be properly updated"
+  end
+  
+  def test_begin_line_without_existing_class
+    # Test begin_line when the span doesn't have a class attribute
+    
+    encoder = CodeRay::Encoders::HTML.new
+    encoder.instance_variable_set(:@out, '')
+    encoder.instance_variable_set(:@opened, [])
+    encoder.instance_variable_set(:@span_for_kinds, {
+      # Span with '>' in attribute but no class
+      :test => '<span title="a > b > c" data-info="x > y">'
+    })
+    
+    encoder.send(:begin_line, :test)
+    result = encoder.instance_variable_get(:@out)
+    
+    # Should add class="line" at the end, preserving all existing attributes
+    assert_equal '<span title="a > b > c" data-info="x > y" class="line">', result
+    assert result.include?('title="a > b > c"'), "Title attribute should remain intact"
+    assert result.include?('data-info="x > y"'), "Data attribute should remain intact"
+    assert result.include?('class="line"'), "Line class should be added"
+  end
+  
+  def test_begin_line_normal_case
+    # Test the normal case without special characters
+    
+    encoder = CodeRay::Encoders::HTML.new
+    encoder.instance_variable_set(:@out, '')
+    encoder.instance_variable_set(:@opened, [])
+    encoder.instance_variable_set(:@span_for_kinds, {
+      :test => '<span class="keyword">'
+    })
+    
+    encoder.send(:begin_line, :test)
+    result = encoder.instance_variable_get(:@out)
+    
+    assert_equal '<span class="line keyword">', result
+  end
+  
+  def test_begin_line_no_style
+    # Test when there's no predefined style for the kind
+    
+    encoder = CodeRay::Encoders::HTML.new
+    encoder.instance_variable_set(:@out, '')
+    encoder.instance_variable_set(:@opened, [])
+    encoder.instance_variable_set(:@span_for_kinds, {})
+    
+    encoder.send(:begin_line, :unknown_kind)
+    result = encoder.instance_variable_get(:@out)
+    
+    assert_equal '<span class="line">', result
+  end
+end
\ No newline at end of file