diff --git a/test/test_transforms.py b/test/test_transforms.py
index 18bed4ebabf..d7739ccbd6a 100644
--- a/test/test_transforms.py
+++ b/test/test_transforms.py
@@ -284,6 +284,20 @@ def test_added_transforms_are_in_eval_mode():
     assert t.transform[1].training
 
 
+def test_nested_transformed_env():
+    base_env = ContinuousActionVecMockEnv()
+    t1 = RewardScaling(0, 1)
+    t2 = RewardScaling(0, 2)
+    env = TransformedEnv(TransformedEnv(base_env, t1), t2)
+
+    assert env.base_env is base_env
+    assert isinstance(env.transform, Compose)
+    children = list(env.transform.transforms.children())
+    assert len(children) == 2
+    assert children[0] == t1
+    assert children[1] == t2
+
+
 class TestTransforms:
     @pytest.mark.skipif(not _has_tv, reason="no torchvision")
     @pytest.mark.parametrize("interpolation", ["bilinear", "bicubic"])
@@ -1320,7 +1334,13 @@ def test_r3m_instantiation(self, model, tensor_pixels_key, device):
         transformed_env.close()
 
     @pytest.mark.parametrize("stack_images", [True, False])
-    @pytest.mark.parametrize("parallel", [True, False])
+    @pytest.mark.parametrize(
+        "parallel",
+        [
+            False,
+            True,
+        ],
+    )
     def test_r3m_mult_images(self, model, device, stack_images, parallel):
         keys_in = ["next_pixels", "next_pixels2"]
         keys_out = ["next_vec"] if stack_images else ["next_vec", "next_vec2"]
diff --git a/torchrl/envs/transforms/transforms.py b/torchrl/envs/transforms/transforms.py
index 350a5bc175b..bb03ac954d4 100644
--- a/torchrl/envs/transforms/transforms.py
+++ b/torchrl/envs/transforms/transforms.py
@@ -244,10 +244,20 @@ def parent(self) -> EnvBase:
                 raise ValueError(
                     "A transform parent must be either another Compose transform or an environment object."
                 )
+            compose = parent
+            # the parent of the compose must be a TransformedEnv
+            compose_parent = compose.parent
+            if not isinstance(compose_parent, TransformedEnv):
+                raise ValueError(
+                    f"Compose parent was of type {type(compose_parent)} but expected TransformedEnv."
+                )
             out = TransformedEnv(
-                parent.parent.base_env,
+                compose_parent.base_env,
+                transform=compose_parent.transform
+                if compose_parent.transform is not compose
+                else None,
             )
-            for transform in parent.transforms:
+            for transform in compose.transforms:
                 if transform is self:
                     break
                 out.append_transform(transform)
@@ -291,21 +301,31 @@ def __init__(
         cache_specs: bool = True,
         **kwargs,
     ):
+        self._transform = None
         device = kwargs.pop("device", env.device)
         env = env.to(device)
         super().__init__(device=None, **kwargs)
-        self._set_env(env, device)
-        if transform is None:
-            transform = Compose()
-            transform.set_parent(self)
+
+        if isinstance(env, TransformedEnv):
+            self._set_env(env.base_env, device)
+            if type(transform) is not Compose:
+                # we don't use isinstance as some transforms may be subclassed from
+                # Compose but with other features that we don't want to loose.
+                transform = [transform]
+            env_transform = env.transform
+            if type(env_transform) is not Compose:
+                env_transform = [env_transform]
+            transform = Compose(*env_transform, *transform).to(device)
         else:
-            transform = transform.to(device)
-        transform.eval()
+            self._set_env(env, device)
+            if transform is None:
+                transform = Compose()
+            else:
+                transform = transform.to(device)
         self.transform = transform
 
         self._last_obs = None
         self.cache_specs = cache_specs
-
         self._reward_spec = None
         self._observation_spec = None
         self.batch_size = self.base_env.batch_size
@@ -315,6 +335,21 @@ def _set_env(self, env: EnvBase, device) -> None:
         # updates need not be inplace, as transforms may modify values out-place
         self.base_env._inplace_update = False
 
+    @property
+    def transform(self) -> Transform:
+        return self._transform
+
+    @transform.setter
+    def transform(self, transform: Transform):
+        if not isinstance(transform, Transform):
+            raise ValueError(
+                f"""Expected a transform of type torchrl.envs.transforms.Transform,
+but got an object of type {type(transform)}."""
+            )
+        transform.set_parent(self)
+        transform.eval()
+        self._transform = transform
+
     @property
     def device(self) -> bool:
         return self.base_env.device