diff --git a/database.py b/database.py
index 6edbe51..ebafd93 100644
--- a/database.py
+++ b/database.py
@@ -13,7 +13,7 @@ schema = dj.schema("fish_book", locals())
 
 @schema
 class Dataset(dj.Manual):
-    definition = """ # Datasets
+    definition = """ # Dataset
        dataset_id : varchar(256)
        ----
        data_source : varchar(512) # path to the dataset
@@ -59,6 +59,71 @@ class Dataset(dj.Manual):
         return sane
 
 
+
+@schema
+class Subject(dj.Manual):
+    definition = """
+    # Subject
+    subject_id : varchar(256)
+    ----
+    species : varchar(256) 
+    """
+
+    @staticmethod
+    def get_template_tuple(subject_id=None):
+        tup = dict(subject_id=None, species="")
+        if subject_id is not None:
+            d = dict((Subject() & {"subject_id": subject_id}).fetch1())
+            return d
+        return tup
+
+    def make(self, key):
+        file_path = Dataset.get_nix_file(key)
+        if file_path is None:
+            return
+        nix_file = nix.File.open(file_path, nix.FileMode.ReadOnly)
+        m = nix_file.blocks[0].metadata
+        inserts = Subject.get_template_tuple()
+        subj_info = m["Recording"]["Subject"]
+        inserts["subject_id"] = subj_info["Identifier"]
+        inserts["species"] = subj_info["Species"][0]
+        inserts["weight"] = subj_info["Weight"]
+        inserts["size"] = subj_info["Size"]
+        inserts["eod_frequency"] = np.round(subj_info["EOD Frequency"] * 10) / 10
+        inserts.update(key)
+        self.insert1(inserts, skip_duplicates=True)
+        nix_file.close()
+
+
+@schema
+class SubjectDatasetMap(dj.Manual):
+    definition = """
+    # SubjectDatasetMap
+    -> Subject
+    -> Dataset
+    """
+
+
+@schema
+class SubjectProperties(dj.Manual):
+    definition = """
+    # SubjectProperties
+    id : int auto_increment
+    ---- 
+    -> Subject
+    recording_date : date
+    weight : float
+    size : float
+    eod_frequency : float
+    """
+
+    def get_template_tuple(id=None):
+        tup = dict(id=None, subject_id=None, recording_date=None, weight=0.0, size=0.0, eod_frequency=0.0)
+        if id is not None:
+            return dict((SubjectProperties() & {"id": id}).fetch1())
+        return tup
+
+
 def read_info(info_file):
     if not os.path.exists(info_file):
         return None, None, False
@@ -94,7 +159,49 @@ def populate_datasets(data_path):
     Dataset().insert1(inserts, skip_duplicates=True)
 
 
+def populate_subjects(data_path):
+    print("Importing subject(s) of %s" % data_path)
+    dset_name = os.path.split(data_path)[-1]
+    info_file = os.path.join(data_path, 'info.dat')
+    if not os.path.exists(info_file):
+        return None, None, False
+    info = read_info_file(info_file)
+    p = []
+    find_key_recursive(info, "Subject", p)
+    if len(p) > 0:
+        subj = deep_get(info, p)
+    inserts = Subject.get_template_tuple()
+    inserts["subject_id"] = subj["Identifier"]
+    inserts["species"] = subj["Species"]
+    Subject().insert1(inserts, skip_duplicates=True)
+
+    # multi mach entry
+    dataset = dict((Dataset() & {"dataset_id": dset_name}).fetch1())
+    mm = dict(dataset_id=dataset["dataset_id"], subject_id=subj["Identifier"])
+    SubjectDatasetMap.insert1(mm, skip_duplicates=True)
+
+    # subject properties
+    props = SubjectProperties.get_template_tuple()
+    props["subject_id"] = subj["Identifier"]
+    props["recording_date"] = dataset["recording_date"]
+    if "Weight" in subj.keys():
+        props["weight"] = np.round(float(subj["Weight"][:-1]), 1)
+    if "Size" in subj.keys():
+        props["size"] = np.round(float(subj["Size"][:-2]), 1)
+    if "EOD Frequency" in subj.keys():
+        props["eod_frequency"] = np.round(float(subj["EOD Frequency"][:-2]))
+    p = props.copy()
+    p.pop("id")
+    if len(SubjectProperties & p) == 0:
+        SubjectProperties.insert1(props, skip_duplicates=True)
+
+
 if __name__ == "__main__":
     datasets = glob.glob('/data/apteronotus/2018-05-08*')
+    # Dataset.drop()
+    # Subject.drop()
+    # SubjectProperties.drop()
+    # SubjectDatasetMap.drop()
     for d in datasets:
         populate_datasets(d)
+        populate_subjects(d)