It turns out, to make SL6 WN log in a format that 
sysklogd-1.4.1-46.el5.x86_64.rpm on SL5 can grok appears easy.

http://serverfault.com/questions/324871/redirecting-syslog-events-from-rhel-6-to-rhel-5-is-it-possible-to-provide-with

"You can customize rsyslog with templates:

$template sysklogd,"<%PRI%>%TIMESTAMP% %syslogtag%%msg%"
*.*     @192.168.1.1;sysklogd
"  (Thanks to that answerer very much!)

Try on sm05 & sm10:
$template sysklogd,"<%PRI%>%TIMESTAMP% %syslogtag%%msg%"
*.* @sm00.phy.bris.ac.uk:514;sysklogd
*.* @smnat.phy.bris.ac.uk:514;sysklogd

Restart rsyslog on sm05 & sm10; on the SL5 central log/mon host,

root@smnat> cat /var/log/messages |/usr/bin/perl /usr/share/logwatch/scripts/shared/onlyservice 'smartd' > /tmp/m; wc /tmp/m 
  8 105 903 /tmp/m
And now
logwatch --debug 6 --detail 5 --service smartd --range Today --print
shows as expected
 
 --------------------- Smartd Begin ------------------------
 
 DEBUG: Inside OnlyService for smartd
 
 Currently unreadable (pending) sectors detected:
        /dev/sda [SAT] - 3 Time(s)
        37 unreadable sectors detected
 
 Offline uncorrectable sectors detected:
        /dev/sda [SAT] - 3 Time(s)
        37 offline uncorrectable sectors detected
 
 ---------------------- Smartd End -------------------------
 
Right! Onto fixing those WN disks.... blast. It was quiet. Too quiet.

Winnie Lacesso / Bristol University Particle Physics Computing Systems
HH Wills Physics Laboratory, Tyndall Avenue, Bristol, BS8 1TL, UK