All, I realize this is not the question John Carmichael originally asked, but I decided to find out how much the Equation of Time varies over several years on the SAME month and day. I used Jean Meeus's Astronomical Algorithms, chapter 27 (actually, the method is attributed to W. M. Smart) and adopted noon in the middle of North America as the test location.
For the years 2000, 2001, 2002, and 2003, here are the dates of greatest and least "spread" in EoT values: Date Ave. EoT "Spread" in EoT values ================================== Feb. 11 -14m 16s 0 sec Mar. 27 -5m 17s 13 sec May 14 +3m 40s 0 sec June 19 -1m 22s 9 sec July 25 -6m 31s 0 sec Sept 17 +5m 34s 16 sec Nov. 2 +16m 28s 0 sec Dec 22 +1m 17s 22 seconds (the max for the year) What this means is that you can incorporate the AVERAGE value for EoT (for example, 1m 17s on December 22nd) in a sundial's design or auxiliary table, and the reading will never be off more than half of 22 seconds, or 11 seconds. That's pretty good! -- Roger -