NavList:

Mark Coady, you wrote:

Great. I'm pleased to hear you tried this out. And I agree, the results are spectacular, and the technique is easy. It is definitely the best method I've found (and I should add that I 'invented' this one) for determining the exact index error of a sextant or, if you prefer, zeroing it out, as in your case.

I haven't described this in a while so maybe some details would help for everyone following along...

You place your sextant on its side on a stable, level table, standing on its legs. Then you remove the standard telescope from the instrument. Next you place a small spotting scope in line with the usual position of the sextant's telescope. The spotting scope should have a magnification of 20x to 40x. It can sit on the table "behind" the sextant, or it can be mounted on a tripod or other support next to the table. A tripod is probably the best choice. The idea here is that you are replacing the sextant's low magnification optics with high magnification. Index error does not depend on magnification. Though there have been occasional reports from experimental NavList readers suggesting that it does depend on magnification, I am convinced that this is a statistical mirage. With the high magnification of a spotting scope, adjustments of a tenth of a minute of arc are readily visible, and it is easy to observe the sextant's index error to a tenth of a minute reliably and repeatably.

This observation should be done in daylight using a distant landmark like a flagpole or radio tower or any sharp point on a building (a lightning rod, for example). Distance matters! Figuring that the instrument parallax distance between the two light paths is about three inches or 0.25 feet on an "average" sextant, we need a ratio of that size to the actual distance to the object that is less than the desired angular precision, a tenth of a minute of arc, expressed as ratio. As always to convert an angle in minutes to a ratio we compare it to that magic number 3438 (actually 180·60/pi, but close enough). So then we need a distance, D, such that

D/(0.25 feet) = 3438/0.1', or

D = 8595 feet = 1.414 nautical miles.

I've given the final distance to three digits past the decimal point only because it's a fun coincidence that it matches the square root of two. Call it 1.5 n.m. for a round number. That's really the minimum distance. For a target at that distance the error in the measured index error would be as large as 0.1 minutes of arc. If you can see a good target object that's five nautical miles away, then the error in the measured index error is really negligible.

The benefit of this method is clear --it's accurate and easy, and it's essential for observations like lunars where we "chase tenths", to borrow Bill Morris' very apt phrase. There's another advantage in this method for a sextant experimenter: you can test out the various anomalies of index error that we've all heard about. For example, does index error change over time? It does, but just how reliable is it from one day to the next? Or from one month to the next? Or, if it's a plastic sextant, from one minute to the next? Similarly, you can experiment with popular techniques like completing the motion on the micrometer drum alternately clockwise and counter-clockwise. Does it really matter? On some sextants you will discover noticeable differences in index correction depending on the final motion. Others, not at all.

Apart from buying or borrowing a spotting scope, this technique for measuring index error requires no other special tools. You can often find picnic tables on the shore or in parks with nice views of distant objects. It can be helpful to bring along some black cardstock and a bit of tape to make a baffle or shield around the horizon mirror so that the only view you see is through the sextant's optics. 

Incidentally, over the past few years I have become a big fan of zeroing out index error rather than measuring it. It used to be said that sextants which were over-adjusted would develop loose fittings. That's probably true, and it's a good argument for measuring the error rather than eliminating it. But in the 21st century sextants are used with much lower frequency than fifty or a hundred years ago. There's a lot of convenience in zero or near zero index correction especially when you may be using several different instruments.

Frank Reed
ReedNavigation.com
Conanicut Island USA