diff --git a/recirq/fermi_hubbard/publication.py b/recirq/fermi_hubbard/publication.py
index 0b87ee39..9987bf1f 100644
--- a/recirq/fermi_hubbard/publication.py
+++ b/recirq/fermi_hubbard/publication.py
@@ -13,10 +13,11 @@
 # limitations under the License.
 """Data specific to experiment published in arXiv:2010.07965."""
 
-from io import BytesIO
 from copy import deepcopy
+from io import BytesIO
 import os
-from typing import Callable, List, Optional, Tuple
+import re
+from typing import Callable, Optional, Tuple
 from urllib.request import urlopen
 from zipfile import ZipFile
 
@@ -217,7 +218,7 @@ def rainbow23_layouts(sites_count: int = 8) -> Tuple[ZigZagLayout]:
 
 def fetch_publication_data(
         base_dir: Optional[str] = None,
-        exclude: Optional[List[str]] = None,
+        exclude: Optional[Tuple[str]] = (),
 ) -> None:
     """Downloads and extracts publication data from the Dryad repository at
     https://doi.org/10.5061/dryad.crjdfn32v, saving to disk.
@@ -239,23 +240,26 @@ def fetch_publication_data(
     if base_dir is None:
         base_dir = "fermi_hubbard_data"
 
-    base_url = "https://datadryad.org/stash/downloads/file_stream/"
-    data = {
-        "gaussians_1u1d_nofloquet": "451326",
-        "gaussians_1u1d": "451327",
-        "trapping_2u2d": "451328",
-        "trapping_3u3d": "451329"
-    }
-    if exclude is not None:
-        data = {path: key for path, key in data.items() if path not in exclude}
-
-    for path, key in data.items():
+    fnames = {
+        "gaussians_1u1d_nofloquet", "gaussians_1u1d", "trapping_2u2d", "trapping_3u3d"
+    }.difference(exclude)
+
+    # Determine file IDs. Note these are not permanent on Dryad.
+    file_ids = {}
+    for line in urlopen("https://doi.org/10.5061/dryad.crjdfn32v").readlines():
+        for fname in fnames:
+            if fname + ".zip" in line.decode():
+                file_id, = re.findall(r"file_stream/[\d]*\d+", line.decode())
+                file_ids.update({fname: file_id})
+
+    # Download and extract files using IDs.
+    for path, key in file_ids.items():
         print(f"Downloading {path}...")
-        if os.path.exists(path=base_dir + os.path.sep + path):
+        if os.path.exists(path=os.path.join(base_dir, path)):
             print("Data already exists.\n")
             continue
 
-        with urlopen(base_url + key) as stream:
+        with urlopen("https://datadryad.org/stash/downloads/" + key) as stream:
             with ZipFile(BytesIO(stream.read())) as zfile:
                 zfile.extractall(base_dir)
 
diff --git a/recirq/fermi_hubbard/publication_test.py b/recirq/fermi_hubbard/publication_test.py
index 48cc481c..bd2a4c7a 100644
--- a/recirq/fermi_hubbard/publication_test.py
+++ b/recirq/fermi_hubbard/publication_test.py
@@ -18,10 +18,10 @@
 
 def test_fetch_publication_data():
     base_dir = "fermi_hubbard_data"
-    fetch_publication_data(base_dir=base_dir, exclude=["trapping_3u3d"])
+    fetch_publication_data(base_dir=base_dir, exclude=("trapping_3u3d",))
 
     for path in ("gaussians_1u1d_nofloquet", "gaussians_1u1d", "trapping_2u2d"):
-        assert os.path.exists(base_dir + os.path.sep + path)
+        assert os.path.exists(os.path.join(base_dir, path))
 
     fetch_publication_data(base_dir=base_dir)
-    assert os.path.exists(base_dir + os.path.sep + "trapping_3u3d")
+    assert os.path.exists(os.path.join(base_dir, "trapping_3u3d"))