diff --git a/lib/iris/_lazy_data.py b/lib/iris/_lazy_data.py
index a23d47a2e0..7d2ae2a3ad 100644
--- a/lib/iris/_lazy_data.py
+++ b/lib/iris/_lazy_data.py
@@ -72,8 +72,8 @@ def _optimum_chunksize_internals(
 
     Args:
 
-    * chunks (tuple of int, or None):
-        Pre-existing chunk shape of the target data : None if unknown.
+    * chunks (tuple of int):
+        Pre-existing chunk shape of the target data.
     * shape (tuple of int):
         The full array shape of the target data.
     * limit (int):
@@ -106,8 +106,6 @@ def _optimum_chunksize_internals(
         "chunks = [c[0] for c in normalise_chunks('auto', ...)]".
 
     """
-    if chunks is None:
-        chunks = list(shape)
 
     # Set the chunksize limit.
     if limit is None:
diff --git a/lib/iris/fileformats/netcdf/loader.py b/lib/iris/fileformats/netcdf/loader.py
index 107415c79d..b7c1c80cf4 100644
--- a/lib/iris/fileformats/netcdf/loader.py
+++ b/lib/iris/fileformats/netcdf/loader.py
@@ -780,7 +780,13 @@ def set(
     @contextmanager
     def from_file(self) -> None:
         """
-        Ensures the chunks are loaded in from file variables, else will throw an error.
+        Ensures the chunksizes are loaded in from NetCDF file variables.
+
+        Raises
+        ------
+        KeyError
+            If any NetCDF data variables - those that become
+            :class:`~iris.cube.Cube`\\ s - do not specify chunksizes.
 
         Notes
         -----
@@ -798,7 +804,7 @@ def from_file(self) -> None:
     @contextmanager
     def as_dask(self) -> None:
         """
-        Ensures the chunks are decided from dask.
+        Uses Dask :external+dask:doc:`array` to control chunksizes.
 
         Notes
         -----