diff --git a/README.md b/README.md
index dce792cc..aac57ed7 100644
--- a/README.md
+++ b/README.md
@@ -1224,7 +1224,7 @@ Selectively reading only the necessary data in this way is particularly useful f
 
 ### Spatial filtering
 
-Another common use-case is that we only want to read those records that are located in some region of interest. Because the shapefile stores the bounding box of each shape separately from the geometry data, it's possible to quickly retrieve all shapes that might overlap a given bounding box region without having to load the full shape geometry data for every shape. This can be done by specifying the `bbox` argument to any of the record or shape methods:
+Another common use-case is that we only want to read those records that are located in some region of interest. Because the shapefile stores the bounding box of each shape separately from the geometry data, it's possible to quickly retrieve all shapes that might overlap a given bounding box region without having to load the full shape geometry data for every shape. This can be done by specifying the `bbox` argument to the shapes, iterShapes, or iterShapeRecords methods:
 
 
 	>>> bbox = [36.423, 12.360, 43.123, 18.004] # ca bbox of Eritrea
diff --git a/shapefile.py b/shapefile.py
index 04fb5ec0..325f284c 100644
--- a/shapefile.py
+++ b/shapefile.py
@@ -1360,6 +1360,13 @@ def __shape(self, oid=None, bbox=None):
         # Read a single point
         if shapeType in (1,11,21):
             record.points = [_Array('d', unpack("<2d", f.read(16)))]
+            if bbox is not None:
+                # create bounding box for Point by duplicating coordinates
+                point_bbox = list(record.points[0] + record.points[0])
+                # skip shape if no overlap with bounding box
+                if not bbox_overlap(bbox, point_bbox):
+                    f.seek(next)
+                    return None
         # Read a single Z value
         if shapeType == 11:
             record.z = list(unpack("<d", f.read(8)))