NavList:

Brad,

Circa 1900, thanks to the language used in various contemporary sources on sextant errors, what we today refer to as "arc error" today in NavList messages was usually listed as "centering error" or (synonymously) "eccentricity". But they measured it just the way we would measure arc error today: check what the instrument reads when it is "fed" collimated light with a series of known angular separations. The table on a sextant certificate, listing offsets every 15 or 30 degrees, might well have been interpreted by sextant experts back then as an eccentricity table, even though the actual origin of the measured offsets was more complex. But it didn't matter.

You added:
"One reference I found seemed to indicate that 'Class A' sextants needed to resolve to at least 10 seconds. That may be true as lesser resolutions were produced."

Again, it is my understanding that a class A sextant had to pass a bunch of tests, some of them would have been simple examinations, like checking the resolution of the scale. You can imagine a junior technician doing a cursory examination on each sextant as it arrives at Kew Obs: index arm moves freely... check box 1... mirrors unbroken... check box 2... arc divided to 10 seconds... check box 3. Other aspects of the sextant, more critical and interesting by our standards today, would have required actual optical testing, like those tests for arc error (centering error). Sextant filters were also tested for prismatic error at Kew, and I would wager that there was some maximum allowed error for that test, too. They would run an instrument through a battery of tests and examinations, and if it passed within narrow bounds, it was "class A". If it failed some tests within some limiting bounds, it was "class B". Still more variances would make it "class C". Apparently those were the only classifications. Any instrument which couldn't meet the minimum standards for a "class C" certificate simply did not receive an official certificate. And as various authors have noted, those uncertified sextants still probably sold well though at lower prices. Also, I strongly suspect that the details of the classifications changed over time. For example, there were undoubtedly tests for micrometers in later decades added to the classification scheme. Kew Observatory ceased its sextant testing operations sometime around 1910 (the details are in a NavList message of mine somewhere), and the task was transferred to another laboratory.

You also wrote:
"What about accuracy of scale, etc? You've got perfect centering and 10 second resolution, but the arc could be divided by a drunken monkey. "

Yes. As I noted above, it's all one problem. Although they called it "centering error", this name could include arc error from any source. A navigator gains nothing from separating the causes. All that counts is the error at each degree (or tenth or thirtieth or whatever) as printed on the table.

-FER

----------------------------------------------------------------
NavList message boards and member settings: www.navlist.net/NavList
Members may optionally receive posts by email.
To cancel email delivery, send a message to NoMail[at]navlist.net
----------------------------------------------------------------