diff --git a/TODO.TXT b/TODO.TXT
index 85081d548e..506cf75a6e 100644
--- a/TODO.TXT
+++ b/TODO.TXT
@@ -37,14 +37,10 @@ that need to be done.  I'm trying to be as granular as possible.
 7)  Remove any references to the global ``django.db.connection`` object in the
     SQL creation process.  This includes(but is probably not limited to):
 
-    * ``django.db.models.sql.where.Where``
-    * ``django.db.models.sql.expressions.SQLEvaluator``
-    * ``django.db.models.sql.query.Query`` uses ``connection`` in place of
-      ``self.connection`` in ``self.add_filter``
     * The way we create ``Query`` from ``BaseQuery`` is awkward and hacky.
     * ``django.db.models.query.delete_objects``
     * ``django.db.models.query.insert_query``
-    * ``django.db.models.base.Model``
+    * ``django.db.models.base.Model`` -- in ``save_base``
     * ``django.db.models.fields.Field`` This uses it, as do it's subclasses.
     * ``django.db.models.fields.related`` It's used all over the place here,
       including opening a cursor and executing queries, so that's going to
@@ -54,6 +50,7 @@ that need to be done.  I'm trying to be as granular as possible.
 
 5)  Add the ``using`` Meta option.  Tests and docs(these are to be assumed at
     each stage from here on out).
+5)  Implement using kwarg on save() method.
 6)  Add the ``using`` method to ``QuerySet``.  This will more or less "just
     work" across multiple databases that use the same backend.  However, it
     will fail gratuitously when trying to use 2 different backends.
diff --git a/django/db/models/base.py b/django/db/models/base.py
index 13ff7e8f35..bb2dbf0504 100644
--- a/django/db/models/base.py
+++ b/django/db/models/base.py
@@ -580,17 +580,16 @@ class Model(object):
     def _get_next_or_previous_in_order(self, is_next):
         cachename = "__%s_order_cache" % is_next
         if not hasattr(self, cachename):
-            qn = connection.ops.quote_name
-            op = is_next and '>' or '<'
+            op = is_next and 'gt' or 'lt'
             order = not is_next and '-_order' or '_order'
             order_field = self._meta.order_with_respect_to
-            # FIXME: When querysets support nested queries, this can be turned
-            # into a pure queryset operation.
-            where = ['%s %s (SELECT %s FROM %s WHERE %s=%%s)' % \
-                (qn('_order'), op, qn('_order'),
-                qn(self._meta.db_table), qn(self._meta.pk.column))]
-            params = [self.pk]
-            obj = self._default_manager.filter(**{order_field.name: getattr(self, order_field.attname)}).extra(where=where, params=params).order_by(order)[:1].get()
+            obj = self._default_manager.filter(**{
+                order_field.name: getattr(self, order_field.attname)
+            }).filter(**{
+                '_order__%s' % op: self._default_manager.values('_order').filter(**{
+                    self._meta.pk.name: self.pk
+                })
+            }).order_by(order)[:1].get()
             setattr(self, cachename, obj)
         return getattr(self, cachename)
 
diff --git a/django/db/models/sql/expressions.py b/django/db/models/sql/expressions.py
index 920cbffe73..826c4f398c 100644
--- a/django/db/models/sql/expressions.py
+++ b/django/db/models/sql/expressions.py
@@ -1,5 +1,4 @@
 from django.core.exceptions import FieldError
-from django.db import connection
 from django.db.models.fields import FieldDoesNotExist
 from django.db.models.sql.constants import LOOKUP_SEP
 
@@ -10,6 +9,7 @@ class SQLEvaluator(object):
         self.cols = {}
 
         self.contains_aggregate = False
+        self.connection = query.connection
         self.expression.prepare(self, query, allow_joins)
 
     def as_sql(self, qn=None):
@@ -19,6 +19,9 @@ class SQLEvaluator(object):
         for node, col in self.cols.items():
             self.cols[node] = (change_map.get(col[0], col[0]), col[1])
 
+    def update_connection(self, connection):
+        self.connection = connection
+
     #####################################################
     # Vistor methods for initial expression preparation #
     #####################################################
@@ -56,7 +59,7 @@ class SQLEvaluator(object):
 
     def evaluate_node(self, node, qn):
         if not qn:
-            qn = connection.ops.quote_name
+            qn = self.connection.ops.quote_name
 
         expressions = []
         expression_params = []
@@ -75,11 +78,11 @@ class SQLEvaluator(object):
                 expressions.append(format % sql)
                 expression_params.extend(params)
 
-        return connection.ops.combine_expression(node.connector, expressions), expression_params
+        return self.connection.ops.combine_expression(node.connector, expressions), expression_params
 
     def evaluate_leaf(self, node, qn):
         if not qn:
-            qn = connection.ops.quote_name
+            qn = self.connection.ops.quote_name
 
         col = self.cols[node]
         if hasattr(col, 'as_sql'):
diff --git a/django/db/models/sql/query.py b/django/db/models/sql/query.py
index d290d60e63..2059139600 100644
--- a/django/db/models/sql/query.py
+++ b/django/db/models/sql/query.py
@@ -67,10 +67,10 @@ class BaseQuery(object):
         # SQL-related attributes
         self.select = []
         self.tables = []    # Aliases in the order they are created.
-        self.where = where()
+        self.where = where(connection=self.connection)
         self.where_class = where
         self.group_by = None
-        self.having = where()
+        self.having = where(connection=self.connection)
         self.order_by = []
         self.low_mark, self.high_mark = 0, None  # Used for offset/limit
         self.distinct = False
@@ -151,6 +151,8 @@ class BaseQuery(object):
         # supported. It's the only class-reference to the module-level
         # connection variable.
         self.connection = connection
+        self.where.update_connection(self.connection)
+        self.having.update_connection(self.connection)
 
     def get_meta(self):
         """
@@ -243,6 +245,8 @@ class BaseQuery(object):
             obj.used_aliases = set()
         obj.filter_is_sticky = False
         obj.__dict__.update(kwargs)
+        obj.where.update_connection(obj.connection) # where and having track their own connection
+        obj.having.update_connection(obj.connection)# we need to keep this up to date
         if hasattr(obj, '_setup_query'):
             obj._setup_query()
         return obj
@@ -530,10 +534,10 @@ class BaseQuery(object):
                 self.where.add(EverythingNode(), AND)
         elif self.where:
             # rhs has an empty where clause.
-            w = self.where_class()
+            w = self.where_class(connection=self.connection)
             w.add(EverythingNode(), AND)
         else:
-            w = self.where_class()
+            w = self.where_class(connection=self.connection)
         self.where.add(w, connector)
 
         # Selection columns and extra extensions are those provided by 'rhs'.
@@ -1534,7 +1538,7 @@ class BaseQuery(object):
             lookup_type = 'isnull'
             value = True
         elif (value == '' and lookup_type == 'exact' and
-              connection.features.interprets_empty_strings_as_nulls):
+              self.connection.features.interprets_empty_strings_as_nulls):
             lookup_type = 'isnull'
             value = True
         elif callable(value):
@@ -1546,7 +1550,7 @@ class BaseQuery(object):
 
         for alias, aggregate in self.aggregates.items():
             if alias == parts[0]:
-                entry = self.where_class()
+                entry = self.where_class(connection=self.connection)
                 entry.add((aggregate, lookup_type, value), AND)
                 if negate:
                     entry.negate()
@@ -1614,7 +1618,7 @@ class BaseQuery(object):
                     for alias in join_list:
                         if self.alias_map[alias][JOIN_TYPE] == self.LOUTER:
                             j_col = self.alias_map[alias][RHS_JOIN_COL]
-                            entry = self.where_class()
+                            entry = self.where_class(connection=self.connection)
                             entry.add((Constraint(alias, j_col, None), 'isnull', True), AND)
                             entry.negate()
                             self.where.add(entry, AND)
@@ -1623,7 +1627,7 @@ class BaseQuery(object):
                     # Leaky abstraction artifact: We have to specifically
                     # exclude the "foo__in=[]" case from this handling, because
                     # it's short-circuited in the Where class.
-                    entry = self.where_class()
+                    entry = self.where_class(connection=self.connection)
                     entry.add((Constraint(alias, col, None), 'isnull', True), AND)
                     entry.negate()
                     self.where.add(entry, AND)
diff --git a/django/db/models/sql/subqueries.py b/django/db/models/sql/subqueries.py
index 0cd393756d..def1ff8ad8 100644
--- a/django/db/models/sql/subqueries.py
+++ b/django/db/models/sql/subqueries.py
@@ -48,7 +48,7 @@ class DeleteQuery(Query):
         for related in cls._meta.get_all_related_many_to_many_objects():
             if not isinstance(related.field, generic.GenericRelation):
                 for offset in range(0, len(pk_list), GET_ITERATOR_CHUNK_SIZE):
-                    where = self.where_class()
+                    where = self.where_class(connection=self.connection)
                     where.add((Constraint(None,
                             related.field.m2m_reverse_name(), related.field),
                             'in',
@@ -57,14 +57,14 @@ class DeleteQuery(Query):
                     self.do_query(related.field.m2m_db_table(), where)
 
         for f in cls._meta.many_to_many:
-            w1 = self.where_class()
+            w1 = self.where_class(connection=self.connection)
             if isinstance(f, generic.GenericRelation):
                 from django.contrib.contenttypes.models import ContentType
                 field = f.rel.to._meta.get_field(f.content_type_field_name)
                 w1.add((Constraint(None, field.column, field), 'exact',
                         ContentType.objects.get_for_model(cls).id), AND)
             for offset in range(0, len(pk_list), GET_ITERATOR_CHUNK_SIZE):
-                where = self.where_class()
+                where = self.where_class(connection=self.connection)
                 where.add((Constraint(None, f.m2m_column_name(), f), 'in',
                         pk_list[offset : offset + GET_ITERATOR_CHUNK_SIZE]),
                         AND)
@@ -81,7 +81,7 @@ class DeleteQuery(Query):
         lot of values in pk_list.
         """
         for offset in range(0, len(pk_list), GET_ITERATOR_CHUNK_SIZE):
-            where = self.where_class()
+            where = self.where_class(connection=self.connection)
             field = self.model._meta.pk
             where.add((Constraint(None, field.column, field), 'in',
                     pk_list[offset : offset + GET_ITERATOR_CHUNK_SIZE]), AND)
@@ -185,7 +185,7 @@ class UpdateQuery(Query):
 
         # Now we adjust the current query: reset the where clause and get rid
         # of all the tables we don't need (since they're in the sub-select).
-        self.where = self.where_class()
+        self.where = self.where_class(connection=self.connection)
         if self.related_updates or must_pre_select:
             # Either we're using the idents in multiple update queries (so
             # don't want them to change), or the db backend doesn't support
@@ -209,7 +209,7 @@ class UpdateQuery(Query):
         This is used by the QuerySet.delete_objects() method.
         """
         for offset in range(0, len(pk_list), GET_ITERATOR_CHUNK_SIZE):
-            self.where = self.where_class()
+            self.where = self.where_class(connection=self.connection)
             f = self.model._meta.pk
             self.where.add((Constraint(None, f.column, f), 'in',
                     pk_list[offset : offset + GET_ITERATOR_CHUNK_SIZE]),
diff --git a/django/db/models/sql/where.py b/django/db/models/sql/where.py
index b27bc992a6..8420e49d95 100644
--- a/django/db/models/sql/where.py
+++ b/django/db/models/sql/where.py
@@ -4,7 +4,6 @@ Code to manage the creation and SQL rendering of 'where' constraints.
 import datetime
 
 from django.utils import tree
-from django.db import connection
 from django.db.models.fields import Field
 from django.db.models.query_utils import QueryWrapper
 from datastructures import EmptyResultSet, FullResultSet
@@ -34,6 +33,18 @@ class WhereNode(tree.Node):
     """
     default = AND
 
+    def __init__(self, *args, **kwargs):
+        self.connection = kwargs.pop('connection', None)
+        super(WhereNode, self).__init__(*args, **kwargs)
+
+    def __getstate__(self):
+        """
+        Don't try to pickle the connection, our Query will restore it for us.
+        """
+        data = self.__dict__.copy()
+        del data['connection']
+        return data
+
     def add(self, data, connector):
         """
         Add a node to the where-tree. If the data is a list or tuple, it is
@@ -53,7 +64,9 @@ class WhereNode(tree.Node):
             value = list(value)
         if hasattr(obj, "process"):
             try:
-                obj, params = obj.process(lookup_type, value)
+                # FIXME We're calling process too early, the connection could
+                # change
+                obj, params = obj.process(lookup_type, value, self.connection)
             except (EmptyShortCircuit, EmptyResultSet):
                 # There are situations where we want to short-circuit any
                 # comparisons and make sure that nothing is returned. One
@@ -78,6 +91,14 @@ class WhereNode(tree.Node):
         super(WhereNode, self).add((obj, lookup_type, annotation, params),
                 connector)
 
+    def update_connection(self, connection):
+        self.connection = connection
+        for child in self.children:
+            if hasattr(child, 'update_connection'):
+                child.update_connection(connection)
+            elif hasattr(child[3], 'update_connection'):
+                child[3].update_connection(connection)
+
     def as_sql(self, qn=None):
         """
         Returns the SQL version of the where clause and the value to be
@@ -88,7 +109,7 @@ class WhereNode(tree.Node):
         recursion).
         """
         if not qn:
-            qn = connection.ops.quote_name
+            qn = self.connection.ops.quote_name
         if not self.children:
             return None, []
         result = []
@@ -153,7 +174,7 @@ class WhereNode(tree.Node):
             field_sql = lvalue.as_sql(quote_func=qn)
 
         if value_annot is datetime.datetime:
-            cast_sql = connection.ops.datetime_cast_sql()
+            cast_sql = self.connection.ops.datetime_cast_sql()
         else:
             cast_sql = '%s'
 
@@ -163,10 +184,10 @@ class WhereNode(tree.Node):
         else:
             extra = ''
 
-        if lookup_type in connection.operators:
-            format = "%s %%s %%s" % (connection.ops.lookup_cast(lookup_type),)
+        if lookup_type in self.connection.operators:
+            format = "%s %%s %%s" % (self.connection.ops.lookup_cast(lookup_type),)
             return (format % (field_sql,
-                              connection.operators[lookup_type] % cast_sql,
+                              self.connection.operators[lookup_type] % cast_sql,
                               extra), params)
 
         if lookup_type == 'in':
@@ -179,15 +200,15 @@ class WhereNode(tree.Node):
         elif lookup_type in ('range', 'year'):
             return ('%s BETWEEN %%s and %%s' % field_sql, params)
         elif lookup_type in ('month', 'day', 'week_day'):
-            return ('%s = %%s' % connection.ops.date_extract_sql(lookup_type, field_sql),
+            return ('%s = %%s' % self.connection.ops.date_extract_sql(lookup_type, field_sql),
                     params)
         elif lookup_type == 'isnull':
             return ('%s IS %sNULL' % (field_sql,
                 (not value_annot and 'NOT ' or '')), ())
         elif lookup_type == 'search':
-            return (connection.ops.fulltext_search_sql(field_sql), params)
+            return (self.connection.ops.fulltext_search_sql(field_sql), params)
         elif lookup_type in ('regex', 'iregex'):
-            return connection.ops.regex_lookup(lookup_type) % (field_sql, cast_sql), params
+            return self.connection.ops.regex_lookup(lookup_type) % (field_sql, cast_sql), params
 
         raise TypeError('Invalid lookup_type: %r' % lookup_type)
 
@@ -202,7 +223,7 @@ class WhereNode(tree.Node):
             lhs = '%s.%s' % (qn(table_alias), qn(name))
         else:
             lhs = qn(name)
-        return connection.ops.field_cast_sql(db_type) % lhs
+        return self.connection.ops.field_cast_sql(db_type) % lhs
 
     def relabel_aliases(self, change_map, node=None):
         """
@@ -257,7 +278,7 @@ class Constraint(object):
     def __init__(self, alias, col, field):
         self.alias, self.col, self.field = alias, col, field
 
-    def process(self, lookup_type, value):
+    def process(self, lookup_type, value, connection):
         """
         Returns a tuple of data suitable for inclusion in a WhereNode
         instance.