NavList:

Just a little more on my earlier post concerning changes in azimuth with distance between assumed and actual position, further examination of AP3270/HO249 – Vol 1 shows, and I should have spotted sooner, that for a N-S position difference, azimuth line error will be zero for a N-S star and maximum for an E-W star.  The opposite will be true for an E-W position difference.  Add to this the high star low star effect and finding the worst case in the publication is difficult, but the greatest I found was two degrees, and I can't find it again now.  Perhaps this is/was one of the constraints on the authors when selecting the seven star batches.  Here's a point for discusion though.  In Marc St Hilaire method, which has the worst effect, slightly incorrect azimuth line, if any, or straight line for curve?

Robin, P182 of the Nautical Almanac does say that if d is greater than 20nm, do a further iteration.  This is interesting, because AP3270/HO248 – Vol 1 used by air navigators, and many yachtsmen until the arrival of programmable calculators, lists values for only whole degrees of latitude, so the N-S difference between assumed and DR position alone can be up to 30nm.  Also, in air navigation where the aim is to get a fix on the chart as quickly as possible after the time already taken up by the use of averaging mechanisms, iterations just wouldn’t be possible, except perhaps by a second navigator with a second chart on a long oceanic flight.  Otherwise, it would take too long (maybe 30 minutes after fix time, by which time you might have flown 240 miles) and you’d get a right crows nest of lines on your chart.  DaveP