diff --git a/skimage/graph/graph_merge.py b/skimage/graph/graph_merge.py
index 379f11fc..69aa94f1 100644
--- a/skimage/graph/graph_merge.py
+++ b/skimage/graph/graph_merge.py
@@ -2,18 +2,62 @@ import numpy as np
 
 
 def _hmerge(rag, x, y, n):
+    """Callback to handle merging nodes by recomputing mean color.
+
+    The method expects that the mean color of `y` is already computed.
+
+    Parameters
+    ----------
+    graph : RAG
+        The graph under consideration.
+    src, dst : int
+        The verices in `graph` to be merged.
+    n : int
+        A neighbor of `src` or `dst` or both.
+
+    Returns
+    -------
+    weight : float
+        The absolute difference of the mean color between node `y` and `n`.
+    """
     diff = rag.node[y]['mean color'] - rag.node[n]['mean color']
     diff = np.linalg.norm(diff)
     return diff
 
 
-def merge_hierarchical(rag, labels, thresh):
+def merge_hierarchical(labels, rag, thresh):
+    """Perform hierarchical merging of a RAG.
 
+    Given an image's labels and its RAG, the method merges the similar nodes
+    until the weight between every two nodes is more than `thresh`.
+
+    Parameters
+    ----------
+    labels : ndarray
+        The array of labels.
+    rag : RAG
+        The Region Adjacency Graph.
+    thresh : float
+        The threshold. Nodes are merged until the minimum edge weight in the
+        graph exceeds `thresh`.
+
+    Returns
+    -------
+    out : ndarray
+        The new labelled array.
+
+    Examples
+    --------
+    >>> from skimage import data, graph, segmentation
+    >>> img = data.coffee()
+    >>> labels = segmentation.slic(img)
+    >>> rag = graph.rag_mean_color(img, labels)
+    >>> new_labels = graph.merge_hierarchical(labels, rag, 40)
+    """
     min_wt = 0
     while min_wt < thresh:
 
-        valid_edges = ((x, y, d)
-                       for x, y, d in rag.edges(data=True) if x != y)
+        valid_edges = ((x, y, d) for x, y, d in rag.edges(data=True) if x != y)
         x, y, d = min(valid_edges, key=lambda x: x[2]['weight'])
         min_wt = d['weight']