Thanks Marty for the above post. If we assume the Navy purchased sextants in batches, there would be no compelling reason to test them in order of manufacturer's serial number. However, the grouping of NO numbers around a particular set of manufacturer's numbers (the bounds of the batch) seems significant. Let's hope a few more sextants show up in the desired range.
We can make the case that we have noted Brandis serial numbers running from 1844 to 5760.
We may make a reasonable surmise that, once upon a time, there was a Brandis numbered 3500.
We have noted N.O. numbers from 34 to at least 9746 (there are some dashed numbers, such as 5083-44 that represent a change in the numbering system, I believe).
We may make a reasonable surmise that, once upon a time, N.O. 1547 was assigned to an instrument.
In the absence of actually finding the long-sought logbooks, we can't say that Brandis 3500
was N.O. 1547. At best, we can say it might have been.
If we do find a Brandis 3500 with a different N.O. number, or a N.O. 1547 that is not serial number 3500, then that would destroy the idea that the box found on Niku may once have held a Navy instrument.