diff --git a/magit.el b/magit.el
index 0de0bfcf..3af0ad3b 100644
--- a/magit.el
+++ b/magit.el
@@ -3045,7 +3045,15 @@ typing and automatically refreshes the status buffer."
     (if (and (not branch-remote)
 	     (not current-prefix-arg))
 	(magit-set push-remote "branch" branch "remote"))
-    (magit-run-git-async "push" "-v" push-remote (format "%s:%s" branch ref-branch))))
+    (magit-run-git-async "push" "-v" push-remote
+                         (if ref-branch
+                             (format "%s:%s" branch ref-branch)
+                           branch))
+    ;; Although git will automatically set up the remote,
+    ;; it doesn't set up the branch to merge (at least as of Git 1.6.6.1),
+    ;; so we have to do that manually.
+    (unless ref-branch
+      (magit-set (concat "refs/heads/" branch) "branch" branch "merge"))))
 
 ;;; Log edit mode
 
diff --git a/magit.texi b/magit.texi
index 5955aa6b..23c8d528 100644
--- a/magit.texi
+++ b/magit.texi
@@ -649,10 +649,12 @@ it, like from any other diff.
 
 Magit will run @code{git push} when you type @kbd{P}.  If you give a
 prefix argument to @kbd{P}, you will be prompted for the repository to
-push to.  When no default remote repositor has been configured yet for
+push to.  When no default remote repository has been configured yet for
 the current branch, you will be prompted as well.  Typing @kbd{P} will
-only push the current branch to the remote.  In other words, it will
-run @code{git push <remote> <branch>}.
+only push the current branch to the remote.  In other words, it will run
+@code{git push <remote> <branch>}.  The branch will be created in the
+remote if it doesn't exist already.  The local branch will be configured
+so that it pulls from the new remote branch.
 
 Typing @kbd{f} will run @code{git remote update}.  With a prefix arg, it
 will prompt for the name of the remote to update.  Typing @kbd{F} will