diff --git a/scikits/learn/feature_extraction/sparse/text.py b/scikits/learn/feature_extraction/sparse/text.py
index 290f4d9f4b17f050c684cb9267d0b77b922c770d..dcc69bcfbb003b0def4bcb9dc0985ecccc5a2ef9 100644
--- a/scikits/learn/feature_extraction/sparse/text.py
+++ b/scikits/learn/feature_extraction/sparse/text.py
@@ -64,13 +64,6 @@ class TfidfTransformer(BaseTfidfTransformer):
             d.setdiag(self.idf)
             X = X * d
 
-        if self.normalize:
-            norms = X.multiply(X).sum(axis=1)
-            norms = np.sqrt(np.array(norms).ravel())
-
-            for doc, token in zip(*X.nonzero()):
-                X[doc, token] /= norms[doc]
-
         return X
 
 class Vectorizer(BaseVectorizer):
@@ -83,10 +76,9 @@ class Vectorizer(BaseVectorizer):
     def __init__(self,
                  analyzer=DEFAULT_ANALYZER,
                  use_tf=True,
-                 use_idf=True,
-                 normalize=False):
+                 use_idf=True):
         self.tc = CountVectorizer(analyzer, dtype=np.float64)
-        self.tfidf = TfidfTransformer(use_tf, use_idf, normalize)
+        self.tfidf = TfidfTransformer(use_tf, use_idf)
 
 class HashingVectorizer(object):
     """Compute term freq vectors using hashed term space in a sparse matrix
diff --git a/scikits/learn/feature_extraction/tests/test_text.py b/scikits/learn/feature_extraction/tests/test_text.py
index 2ce2e325b1ab5ee7855968c961bd29010a1bb19e..6a10c80f471488d0a7058aef40fc666050ba1bae 100644
--- a/scikits/learn/feature_extraction/tests/test_text.py
+++ b/scikits/learn/feature_extraction/tests/test_text.py
@@ -217,11 +217,6 @@ def _test_vectorizer(cv_class, tf_class, v_class):
     assert_array_almost_equal(np.sum(tf, axis=1),
                               [1.0] * n_train)
 
-    # test normalization
-    t3 = tf_class(normalize=True)
-    tfidf_n = toarray(t3.fit(counts_train).transform(counts_train))
-    assert_equal(la.norm(tfidf_n[0]), 1.0)
-
     # test the direct tfidf vectorizer
     # (equivalent to term count vectorizer + tfidf transformer)
     train_data = iter(JUNK_FOOD_DOCS[:-1])
diff --git a/scikits/learn/feature_extraction/text.py b/scikits/learn/feature_extraction/text.py
index 26a7c81ff8c44c2b0cd63a61f62798104f2c68fa..418dcdc8d367fdf5785e9885fda4be705223b9d5 100644
--- a/scikits/learn/feature_extraction/text.py
+++ b/scikits/learn/feature_extraction/text.py
@@ -313,15 +313,11 @@ class BaseTfidfTransformer(BaseEstimator):
 
     use_idf: boolean
         enable inverse-document-frequency reweighting
-
-    normalize: boolean
-        normalize vectors to unit-length
     """
 
-    def __init__(self, use_tf=True, use_idf=True, normalize=False):
+    def __init__(self, use_tf=True, use_idf=True):
         self.use_tf = use_tf
         self.use_idf = use_idf
-        self.normalize = normalize
         self.idf = None
 
 class TfidfTransformer(BaseTfidfTransformer):
@@ -366,9 +362,6 @@ class TfidfTransformer(BaseTfidfTransformer):
         if self.use_idf:
             X *= self.idf
 
-        if self.normalize:
-            X /= np.sqrt(np.sum(X ** 2, axis=1))[:,np.newaxis]
-
         return X
 
 class BaseVectorizer(BaseEstimator):
@@ -428,10 +421,9 @@ class Vectorizer(BaseVectorizer):
     def __init__(self,
                  analyzer=DEFAULT_ANALYZER,
                  use_tf=True,
-                 use_idf=True,
-                 normalize=False):
+                 use_idf=True):
         self.tc = CountVectorizer(analyzer, dtype=np.float64)
-        self.tfidf = TfidfTransformer(use_tf, use_idf, normalize)
+        self.tfidf = TfidfTransformer(use_tf, use_idf)
 
 
 class HashingVectorizer(object):