NavList:

Frank,
Yes, Roswall's text, as well as the other authors that I referred to, is in Swedish (albeit 200 years old Swedish). Roswall also describes lunar observations further on in his book. Here he recommends Borda's reflexion circle to be the best instrument for lunars. He also mentions the circle "recently invented" by Mendozo Rios. When determining index error for this circle he mentions measuring the horizon or a star upon itself, not measurement of the sun's diameter.

I have no answer on when the "nearly obsessive attention to the IE vs IC definition" emerged, but guess it is connected with the introduction of calculators and computers for reducing sights. When writing algorithms it is important that the used variables are clearly defined.
Lars