NavList:

Not much time right now for matters navigational, but couldn't resist having a quick look at an amplitude.  Via a somewhat crude method, using tables for another purpose, have calculated amplitude at between 17d and 18d.

Since at sunset azimuth = 270d + amplitude N, - amplitude S, the derived azimuth becomes about 287.5d.

 

Within half a degree so good enough for most compasses. :-)

My text advises:
"To observe the altitude close to zero, the Sun should be observed when the lower limb lies about a semi-diameter above the horizon, and the Moon when the upper limb touches the horizon - at the time of appearing and disappearing"

 

I was taught that the sun is on the celestial horizon (to avoid another tabular correction) when the lower limb is 2/3 of the sun's diameter above the visible horizon.  I think your text is fairly close in practical use to the adage I learned.  For stars and planets it is when the body is again 2/3 of the _sun's_ diameter above the horizon, so there is a bit more guess work here. 

 

 The moon is altogether different and very difficult to shoot when on the celestial horizon, which is at the time when the upper limb is on the visible horizon.  Now you have to pre-work the amplitude (in the AM) and hope you catch it as it peeks above the horizon.  I think it would be akin of shooting the bearing of a green flash.  Obviously a PM amp is better to shoot, but still hard to get an accurate reading since it will need to be a very quick affair.  A far better method of shooting the moon amplitude is to shoot when the center of the body is on the visible horizon and then make the aforementioned correction _in the opposite direction_ of the sun correction on the visible horizon.  I have yet to shoot a moon amplitude, so can't claim the accuracy of this method.

 

Jeremy