diff --git a/lib/linguist/blob_helper.rb b/lib/linguist/blob_helper.rb
index 76eab7ea..67f6eef3 100644
--- a/lib/linguist/blob_helper.rb
+++ b/lib/linguist/blob_helper.rb
@@ -256,10 +256,16 @@ module Linguist
           # without changing the encoding of `data`, and
           # also--importantly--without having to duplicate many (potentially
           # large) strings.
-          encoded_newlines = ["\r\n", "\r", "\n"].
-            map { |nl| nl.encode(encoding).force_encoding(data.encoding) }
+          begin
+            encoded_newlines = ["\r\n", "\r", "\n"].
+              map { |nl| nl.encode(encoding, "ASCII-8BIT").force_encoding(data.encoding) }
 
-          data.split(Regexp.union(encoded_newlines), -1)
+            data.split(Regexp.union(encoded_newlines), -1)
+          rescue Encoding::ConverterNotFoundError
+            # The data is not splittable in the detected encoding.  Assume it's
+            # one big line.
+            [data]
+          end
         else
           []
         end
diff --git a/samples/Text/iso8859-8-i.txt b/samples/Text/iso8859-8-i.txt
new file mode 100644
index 00000000..ed2bf6c4
--- /dev/null
+++ b/samples/Text/iso8859-8-i.txt
@@ -0,0 +1 @@
+%¿áé
\ No newline at end of file
diff --git a/test/test_blob.rb b/test/test_blob.rb
index da13b96e..2109e9b4 100644
--- a/test/test_blob.rb
+++ b/test/test_blob.rb
@@ -97,6 +97,7 @@ class TestBlob < Test::Unit::TestCase
   def test_sloc
     assert_equal 2, blob("Ruby/foo.rb").sloc
     assert_equal 3, blob("Text/utf16le-windows.txt").sloc
+    assert_equal 1, blob("Text/iso8859-8-i.txt").sloc
   end
 
   def test_encoding