NavList:

The reasons for artificial horizons being used in aircraft sextants is mainly because a useful horizon from an astro point of view is rarely visible because of mountain ranges, cloud, and flying in darkness.  There’s also dip, which needs to be allowed for if using a marine sextant, but this turns out to be not such a problem as might be expected for a light aircraft flying over the sea on a clear say.  The reason is that dip varies very roughly with the square root of the aircrafts height, so whereas a height error of 50m might cause a sextant error of 12.5nm at sea level, it would only make a 2.3nm error at a nominal 400m, which would be a reasonably safe height to fly in a Gipsy Moth.  At 800m a 50m altimeter error would lead to about a 1.6nm sextant error.  There would be two problems.  One would be obtaining horizon dip tables extending above 50m.  They’re rarely seen.  FC would have to make his own from a formula, and he’d need one whose approximations remained valid up to the heights he would be flying at.  The other would be knowing which pressure value to set on his altimeter sub scale for it to remain reading zero at sea level as he flew into areas of changing pressure so that he knew he was entering his dip table with the correct height.  He could get a good idea from a met forecast obtained before he left.  As he wouldn’t be flying particularly high, it would be reasonable for him to descend to almost just above the sea to check this occasionally en-route.   For my calculations above I used dip in minutes = 1.76 SqRt height in metres, which might not be perfect at the higher altitudes but is close enough to make the point. http://www.erikdeman.de/html/sail040h.htm