diff --git a/checkin_notes b/checkin_notes
index ca71e13704..8341a34ab4 100755
--- a/checkin_notes
+++ b/checkin_notes
@@ -13866,3 +13866,9 @@ David  17 June 2004
 
     client/
         hostinfo_network.C
+
+Rom    June 16 2004
+    - Enable the low-water mark for the work queue again.
+
+    client/
+        cs_scheduler.C
diff --git a/client/cs_scheduler.C b/client/cs_scheduler.C
index d8c5be56d6..beb2f49042 100644
--- a/client/cs_scheduler.C
+++ b/client/cs_scheduler.C
@@ -432,11 +432,7 @@ bool CLIENT_STATE::scheduler_rpc_poll() {
         if (exit_when_idle && contacted_sched_server) {
             should_get_work = false;
         } else {
-            // TODO: the following is a kludge to prevent
-            // repeated work requests, but it defeats the
-            // work_buf_min concept
-            //
-            //below_work_buf_min = (current_work_buf_days() <= global_prefs.work_buf_min_days);
+            below_work_buf_min = (current_work_buf_days() <= global_prefs.work_buf_min_days);
             current_work_buf_days(work_buf_days, nactive_results);
             below_work_buf_min = nactive_results < host_info.p_ncpus;
             should_get_work = below_work_buf_min && some_project_rpc_ok();