diff --git a/igneous/task_creation/skeleton.py b/igneous/task_creation/skeleton.py
index 0c6c3901..b6829c0c 100644
--- a/igneous/task_creation/skeleton.py
+++ b/igneous/task_creation/skeleton.py
@@ -349,7 +349,8 @@ def create_sharded_skeleton_merge_tasks(
     ShardedSkeletonMergeTask(
       layer_path, shard_no, 
       dust_threshold, tick_threshold,
-      max_cable_length=max_cable_length
+      max_cable_length=max_cable_length,
+      spatial_index_db=spatial_index_db,
     )
     for shard_no in shard_labels.keys()
   ]
diff --git a/igneous/tasks/skeleton.py b/igneous/tasks/skeleton.py
index fa41449b..419ad52d 100644
--- a/igneous/tasks/skeleton.py
+++ b/igneous/tasks/skeleton.py
@@ -313,11 +313,11 @@ class ShardedSkeletonMergeTask(RegisteredTask):
   def __init__(
     self, cloudpath, shard_no, 
     dust_threshold=4000, tick_threshold=6000,
-    sqlite_db=None, max_cable_length=None
+    spatial_index_db=None, max_cable_length=None
   ):
     super(ShardedSkeletonMergeTask, self).__init__(
       cloudpath, shard_no,  
-      dust_threshold, tick_threshold, sqlite_db,
+      dust_threshold, tick_threshold, spatial_index_db,
       max_cable_length
     )
     self.progress = False
@@ -326,7 +326,11 @@ def __init__(
   def execute(self):
     # cache is necessary for local computation, but on GCE download is very fast
     # so cache isn't necessary.
-    cv = CloudVolume(self.cloudpath, cache=False, progress=self.progress)
+    cv = CloudVolume(
+      self.cloudpath, 
+      progress=self.progress,
+      spatial_index_db=self.spatial_index_db
+    )
 
     # This looks messy because we are trying to avoid retaining
     # unnecessary memory. In the original iteration, this was 
@@ -432,8 +436,6 @@ def get_unfused(self, labels, filenames, cv):
 
   def locations_for_labels(self, labels, cv):
     SPATIAL_EXT = re.compile(r'\.spatial$')
-    if self.sqlite_db:
-      cv.skeleton.spatial_index.sqlite_db = self.sqlite_db
     index_filenames = cv.skeleton.spatial_index.file_locations_per_label(labels)
     for label, locations in index_filenames.items():
       for i, location in enumerate(locations):