Thursday, March 12, 2015

What do the Numbers 3500 and 1542 Tell Us? Part Two

In my first post on topic of the Nikumaroro sextant box numbers, I showed that a good argument can be made that the two numbers the sextant box were marked with, 3500 and 1542, are the respective Brandis and U.S. Naval Observatory serial numbers of the sextant this box once held.  In this post, I will use a few basic assumptions and various kinds of evidence to arrive at a manufacture date for the hypothetical Brandis sextant #3500.  This manufacture date will provide a test of the hypothesis that the Nikumaroro sextant box was made to hold a Brandis sextant: for this to be true, the sextant would have to have been made sometime before the sextant box was found in Nikumaroro in 1939.  The manufacture date also will provide a test of Tighar’s hypothesis that the sextant box arrived at Nikumaroro in Amelia Earhart’s Lockheed Electra: for that to be true, the sextant would have had to have been manufactured sometime before Earhart’s ill-fated 1937 flight.  If the arrived-at manufacture date precedes both of these events, this fact alone would not make these hypotheses true, it would simply indicate that they are plausible.  Other ideas have been put forward about what the Nikumaroro sextant box was and how it arrived at Nikumaroro, and they may also be consistent with the arrived-at manufacture date. 

Collectors of antique precision instruments made by certain companies have at their disposal serial number chronologies that make it possible to determine the date of manufacture of an instrument by its serial number (1).  Unfortunately, a Brandis serial number chronology does not seem to exist, so here I will develop one that provides sufficient detail to place fairly narrow limits on when Brandis manufactured the instrument they assigned serial number 3500 to.  Two general assumptions are needed to develop this chronology.  The first is that Brandis used a single serial number series to identify the various kinds of instruments it made, as was the case for a number of other precision instrument manufacturing companies.  The other is that higher Brandis serial numbers indicate later dates of manufacture; perhaps at a short enough time scale (e.g., days or weeks), this assumption doesn’t apply, but I find it reasonable to think it holds well enough to narrow down an instrument’s manufacture date to within a 1-year time period, and that is sufficient for our purposes here.

The approach I will take here is to establish reference points in the Brandis serial number chronology by identifying the dates of manufacture of a number of reference instruments whose Brandis serial numbers are known; if the above two general assumptions are valid, then instruments whose Brandis serial numbers fall between those of any two reference instruments will have been made sometime between the manufacture dates of those two reference instruments.  Since we only really care about when the hypothetical Brandis #3500 sextant was made, it would be sufficient to identify two reference points that reasonably well constrain its date of manufacture, but I will present a considerably longer chronology, if only to show that various sources of information I’ve found, and the assumptions I’ve made, come together to form a self-consistent Brandis serial number chronology.

The first reference point in the Brandis serial number chronology is provided by a type of land surveying instrument called a transit that had Brandis serial number 1659, and that belonged to the Department of Parks of the Borough of Brooklyn, New York City.  In 1899, this transit was sent to Brandis for repair; a dated bill for this repair (2) is shown below .  The billing date, September 18, 1899, provides a ‘no later than’ date for the manufacture of this instrument; we don’t know when it was made, but we can be sure that it was made before September 18, 1899.

Photo Courtesy of The Biggert Collection of Architectural Vignettes on Commercial Stationery,
Avery Architectural and Fine Arts Library, Columbia University. 

 A sextant with Brandis serial number 1794 is listed among the items that were displayed at the Paris International Exposition of 1900, held from April to November of 1900 (3).  It is reasonable to think that Brandis supplied a recently-made sextant to this exposition, and therefore that this sextant was made in 1899 or early 1900.

A sextant with Brandis serial number 2763 and U.S. Naval Observatory serial number 348 was sold on eBay several years ago (see supporting materials page).  The label on the interior of this sextant’s box is hard to read in the pictures posted on eBay, but according to the seller, the label indicates that the sextant was ordered from Brandis in March of 1911 and delivered to the purchaser on May of the same year; it is reasonable to think this sextant was made in 1911.

The Surveying Antiques web site displays a photo of a case label for a Brandis level (4), a type of surveying instrument, whose Brandis serial number was 2928.  The case label indicates that this level was ordered from Brandis on May 13, 1915 and delivered to the purchaser on June 5, 1915; it is reasonable to think this instrument was made in 1915.

Case Label, Brandis Level # 2928 (Survey Antiques web site)

An article in a 1919 issue of Popular Astronomy (5) presents the findings of a party of U.S. Naval Observatory astronomers who observed a solar eclipse at Mt. Baker, Oregon in June of 1918.  The article mentions that the Naval Observatory astronomers had with them a sextant with Brandis serial number 3257; this instrument must have been made some time before  June, 1918.

Mount Baker Eclipse Party Article, 1919

 So far, dates of manufacture have been established for five Brandis instruments.  The dates of manufacture of the middle three instruments are well-constrained, i.e., the date range specified is pretty narrow, and it’s hard to believe an of them could be off by more than a year.  The ‘no later than’ dates for the first and fifth instruments are less exact, but they nevertheless fit in with the manufacture dates of the middle three instruments, i.e., the ‘no later than’ date for Brandis transit #1659 pre-dates the manufacture date for Brandis sextant #1794, and the ‘no later than’ date for Brandis sextant #3257 post-dates the manufacture date of Brandis level #2928.  So far, we have a self-consistent Brandis serial number chronology with five reasonably well-established time points, but we’re still short of the serial number we are really interested in — Brandis #3500.

To extend the chronology to higher serial numbers, dates of manufacture will be attached to several sextants based on the fact that Brandis changed its name several times during the course of its existence.  A web page of the Smithsonian National Museum of American History (6) provides the following brief history of the instrument maker F. E. Brandis and his company:

“Frederick Ernest Brandis (1845–1916) was born in Germany, came to the United States in 1858, worked for Stackpole & Brother for a few years, and opened his own instrument shop in 1871. The firm became F. Brandis & Co. in 1875, F. E. Brandis, Sons & Co. in 1890, and Brandis & Sons, Inc. in 1916. The Pioneer Instrument Company purchased control of the Brandis firm in 1922, and was in turn acquired by the Bendix Aviation Corporation in 1928. The manufacture of Brandis instruments ceased in 1932.
Ref: Charles E. Smart, The Makers of Surveying Instruments in America Since 1700 (Troy, N.Y., 1962), p. 14-15.”

Brandis stamped its name into the instruments it manufactured, and it is reasonable to assume that the wording of the firm name Brandis stamped into its instruments changed whenever the wording of the Brandis firm name did.  The Smithsonian has used Brandis firm names to estimate the dates of manufacture of instruments in its collection, and I will use this approach as well.  Based on the Charles Smart book that the Smithsonian cites, one would assume that an instrument marked “F. Brandis & Co.” was made no later than 1890, an instrument marked “F. E. Brandis, Sons & Co” was made between 1890 and 1916, and an instrument marked “Brandis & Sons, Inc.” was made in 1916 or later.  I think the Charles Smart book may be wrong about certain aspects of the history of Brandis firm name changes, however, so I will used a somewhat different history of Brandis firm name changes to estimate dates of manufacture of Brandis instruments.  My revisions are pretty minor though, so if Charles Smart has it right and I’ve got it wrong, I’m not introducing any serious errors into the Brandis serial number chronology by using my revised history rather than Charles Smart’s.

In searching through descriptions and photos of Brandis instruments in museum collections, auction houses, eBay, and other sources to determine the specific wording of their Brandis firm name markings, I’ve found four instruments that are marked ‘Brandis & Sons Mfg. Co.’,  a version of the Brandis firm name that the Charles Smart book doesn’t mention (7).  The Brandis level bearing serial number 2928 that was discussed above is one of these four instruments, as its case label indicates .  The other three instruments, all sextants, have the next three highest Brandis serial numbers — 3193, 3239, and 3243 — whose Brandis name markings I’ve been able to read (see the supporting materials page).  The Brandis serial numbers of this group of four instruments are bracketed by the Brandis #2763 discussed above, which is marked ‘FE Brandis, Sons & Co.’, and a sextant with Brandis serial number 3331, which is marked ‘Brandis & Sons, Inc.’ (see photo below).  Thus it appears that there was a period when Brandis was called ‘Brandis and Sons Manufacturing Company’, and that this period came after the ‘FE Brandis & Sons & Company’ era, and before the ‘Brandis & Sons, Inc’ era.

When did these firm name changes occur?  The change from ‘FE Brandis & Sons & Co.’ to ‘Brandis & Sons Mfg. Co.’ must have occurred between 1912 and 1914, because the 1912 edition of the Brooklyn City Directory (8) has Brandis listed as ‘Brandis, FE & Sons & Co Engineering Insts’, while The Handbook of Construction Plants, its Cost and Efficiency Cost, published in 1914 (9), has Brandis listed as ‘Brandis & Sons Mfg. Co. Brooklyn, N.Y.’.  As for the name change from ‘Brandis & Sons Mfg. Co’ to ‘Brandis & Sons, Inc.’, the Charles Smart book gives 1916 as the year that ‘Brandis & Sons, Incorporated’ came into existence, however, in an advertisement in the June 28, 1917 issue of a journal called Engineering News - Record (10) gives the Brandis firm name as 'Brandis & Sons Mfg Company', so it appears that the change to Brandis & Sons, Inc. occurred no earlier than July, 1917.  But the name change occurred before the end of 1917, because an advertisement in the December, 1919 issue of a boating magazine called The Rudder (11) contains an advertisement for Brandis sextants (see below) that includes the words “Incorporated 1917” near the bottom of the advertisement.  It may seem imprudent to prefer information provided by these advertisements to the Charles Smart book, but since Smart’s book missed the ‘Brandis & Sons Mfg. Co.’ era, I will rely on what the two Brandis advertisements suggest and assume that instruments marked 'Brandis & Sons, Inc' were made no earlier than July, 1917.

Engineering News-Record, June 28, 1917; page 140. (Google Books)

The Rudder, December 1919. (Google Books)

The sextants with Brandis serial numbers 3193, 3239, and 3243 must have made sometime between 1915 and 1917, because they have higher, and thus later, serial numbers than the Brandis level # 2928 that was made in 1915, and they bear the ‘Brandis & Sons Mfg. Co.’ firm name, and thus were made before the end 1917, the latest date at which the firm name could have changed to ‘Brandis & Sons, Inc’.

The final time point that I can establish for an instrument with a serial number less than 3500 is provided by a sextant with Brandis serial number 3331 which bears the ‘Brandis & Sons, Inc.’ firm name, and thus could have been made no earlier than July, 1917; this sextant has the lowest Brandis serial number of any instrument I know of bearing this firm name marking (12).  I should say that not every sextant I’ve seen with a Brandis serial number higher than 3331 is marked ‘Brandis & Sons, Inc.’; some are instead marked ‘Brandis and Sons', but the two markings are intermixed with respect to serial number as far as I can tell so I think the latter name marking is simply a shortened version of the former

Brandis Sextant #3331, Brandis & Sons, Inc.  (Land & Sea Collection web site)

The Brandis serial number chronology can be further extended thanks to the U.S. Naval Observatory’s role in the early decades of the 20th century as what we would now call a quality assurance laboratory for the U.S. Navy.  As discussed in Part One of this series of posts on the Nikumaroro sextant box numbers, newly purchased U. S. Navy. sextants were inspected for suitability at the Naval Observatory before they were put into service.  Due to slight manufacturing imperfections, angles measured with many sextants deviate from true angles by small but navigationally significant amounts.  These deviations must be corrected for in order to obtain the most accurate results, therefore part of the Naval Observatory’s inspection process involved determining the necessary corrections.  Attached to the inside of a sextant’s box was an Naval Observatory inspection certificate that listed these corrections for several different angles (see photo below); the inspection certificate also provided spaces to indicate the sextant’s Naval Observatory serial number, the sextant manufacturer’s name and serial number, and the inspection date, although sometimes not all this information would actually be filled in on the certificate.  The inspection date is the key piece of information for our purposes because it provides us with a ‘no later than’ date; the sextant’s manufacture date must have preceded its inspection date (13).

A Naval Observatory Inspection Certificate (from Collectors Weekly web site)

Three Brandis sextants that have survived to the present time have inspection certificates that provide useful time points for the Brandis serial number chronology.  Two of them are standard maritime sextants that were modified into Byrd bubble sextants for use by a group of three U.S. Navy-Curtiss (NC) flying boats that in 1919 attempted to make the first successful flight across the Atlantic Ocean (14); only one flying boat made it all the way across.  One of these two sextants is in the navigational instrument collection of the Smithsonian Institution (15); its Brandis serial number is 5296, USNO serial number 2977, and its inspection certificate is dated March 16, 1919.  The other sextant, which is privately owned (16), has Brandis serial number 5292, USNO serial number 2975, and its inspection certificate is dated March 26, 1919 (the certificate for this sextant is the one shown above).  The third sextant is another instrument in the Smithsonian’s navigational instrument collection; its Brandis serial number is #5620, its USNO serial number is 2939, and its inspection certificate has a 1919 date, but the day and month are not provided by the Smithsonian (17).  The Brandis serial number series appears to have increased far more rapidly between 1917 and 1919 than it did in prior years, and one might speculate that the entry of the United States into World War I in 1917 had something to do with the apparent increase in output of instruments by Brandis; I'll have more to say about this in future posts.

Navy-Curtiss Seaplane NC-4, sometime after the 1919 Translatlantic Flight (U.S. Navy Photo)

I know of only four Brandis instruments with serial numbers higher than 5620, and none of them provide time points that extend the chronology to higher serial numbers.  At the date of this post, Tighar’s sextant list (18) includes over sixty Brandis sextants, only four of which have Brandis serial numbers greater than 5620.  This statistic suggests to me that the four-digit Brandis serial number series ended in the vicinity of Brandis serial number #5953, the highest Brandis serial number that has come to light.

The table below brings together the Brandis serial numbers and manufacture dates of all the instruments that provide useful time points for the Brandis serial number chronology.  Brandis #3500 is placed where it fits in the table, and so I finally arrive at the point that I’ve been laboring toward: according to the chronology developed here: the instrument assigned Brandis serial number 3500 was made sometime between July, 1917 and March, 1919.  From the point of view of when the hypothesized Brandis #3500 sextant would have been made, the hypotheses that the Nikumaroro sextant box once held a Brandis sextant made for the U. S. Navy, and that it arrived on Nikumaroro in Amelia Earhart’s Lockheed Electra, are both reasonable ones.  Of course, alternate hypotheses about what the sextant box is, and how it arrived on Nikumaroro, may also be consistent with this manufacture date.  For instance, some have suggested that the Nikumaroro sextant box came from the freighter Norwich City, which ran aground on Nikumaroro in 1929 (19).  This idea is also consistent with the arrived-at manufacture date, since there would have been 10 to 12 years for the sextant box to make it to the ship’s officer who used it on the Norwich City.

I have additional information that provides a somewhat narrower estimate for when the hypothetical  Brandis #3500/USNO #1542 sextant was made and that provides a check on the Brandis serial number chronology developed here.  However, it would require several more pages to discuss this additional information, and since this has already been a long post I will save that discussion for another day.

Comments, corrections, additional relevant facts, differing viewpoints, etc., are always welcome (no one will be banned, blocked, or castigated for offering differing opinions).  Send to 



 (3) The Report of the Commissioner-General for the United States to the International Universal Exposition, Paris, 1900. Volume III, page 255.  Washington, Government Printing Office, 1901.  Available online at Google Books.


(5) Popular Astronomy, Volume 27, page 359-365; 1919 (month not known).  Accessed at the SAO /NASA Astrophysics Data System:

(6) Smithsonian: [] 

(7)  I’m assuming that the quote above from the Smithsonian is a correct reading of Smart’s book, which I don’t have access to.

(8) The Brooklyn City Directory, Volume LXXXVIII, 1912.  Published by The Brooklyn City Directory Co, 1912.  Digitized by the Internet Archive in 2013.

(9) The Handbook of Construction Plants, its Cost and Efficiency Cost, Richard T. Dana.  Myron C. Clark Publishing Co., Chicago. 1914.  Published in 2013 by Forgotten Books,

(10) Engineering News-Record, June 28, 1917; page 140. Accessed at:

(11) The Rudder, December, 1917.  Accessed at:;seq=644;size=150;view=image


(13) At least some sextants were re-inspected once in service, as proven by examples of sextant boxes with two inspection certificates.  The certificate doesn’t tell us how much time elapsed between the sextant’s manufacture and its first inspection at the Naval Observatory, or whether this was the sextant’s first inspection, thus all it provides us with is a ‘no later than’ date for when the sextant was manufactured and  assigned its Brandis serial number.