Guides

Serials that lie

A serial number rarely lies on purpose. It misleads because of how factories worked, and because the tables collectors use were rebuilt after the fact. These are the patterns that catch people out, each tied to the rule the decoder applies.

The number is not the build date

Martin numbers each instrument as it is finished, and the year assigned to a number is the one in which the highest serial reached that point. A guitar finished in the first weeks of a year can therefore carry a number close to the previous year-end figure. In 1956 the two main published sources disagree on the year-end figure, so numbers between 152775 and 153225 could fall in either 1956 or 1957. The decoder reports that as a doubt, not as a year.

Martin also sold instruments under another name. Serials from 900001 to 902908 are Sigma-Martins from 1981 and 1982, and the Martin table does not apply to them.

Numbers that were reused or run out of order

Gibson’s 1961 to 1972 numbering is the textbook case. Number blocks were reused and used in the wrong order, so a number can fall in more than one year, and a higher number is not always later. The decoder treats a match here as a range of candidates and says so. From 1970 the Norlin era brought six-digit numbers not tied to a year, and the 1975 to 1977 two-digit prefixes are thinly documented. In all these years, pickups, pots, the bridge, the logo and the “Made in USA” stamp are better evidence than the digits.

Fender numbers overlapped too. In the 1950 to 1964 era the numbers were not strictly sequential, and the published bands are broad on purpose. A neck date pencilled on the heel, a pot code and body features all beat the serial.

Digits that repeat

Any scheme with a single year digit repeats every ten years. Gibson’s five-digit system of 1953 to 1961 uses the first digit as the last digit of the year, so a number starting with 8 is 1958. But any five-digit number can also match the 1961 to 1970 ranges, so the digit alone does not settle it. Fender’s N prefix covers 1990 to 2002, and N9 can be 1990 or 1999. MN, for Mexico, has the same problem. Rickenbacker’s 1987 to 1996 scheme repeats its year digit and overlaps the later scheme that starts with the year.

Hyphens change the reading

Gretsch hollowbodies from 1973 to 1981 are stamped month, hyphen, year digit, sequence, such as 3-4127. Enter the hyphen. Without it the same digits fit the 1966 to 1972 date code or the sequential range, and the answer changes by a decade. The decoder prefers hyphenated formats when you type a hyphen, and lists the others as ruled out, with the reason.

Schemes that start and end when sources disagree

Some boundaries are collector consensus, not factory records. The decoder marks those readings “estimated”. Gretsch’s date-code era starts in early 1965 in some sources and mid-1966 in others. Rickenbacker’s letter scheme starts in late 1960 or 1961 depending on the source. Treat estimated edges as soft.

Numbers that carry no year

Harmony, Kay and many Silvertone and Danelectro instruments carry stamped numbers with no published date system. Guild’s model-prefix eras do not map to a year without the model. Pre-1990 Epiphone numbers have documented sequences for parts of the range only. For these, “cannot tell” is the correct answer. Date them from labels, hardware and features instead.

Replaced parts

A serial lives on one part. A neck plate, headstock or jackplate can be swapped, and a replacement neck brings its own number. If the serial sits on a different part from the body, or the finish around it looks newer than the rest, treat the number as describing that part only. Pots and pickups are also changed in repairs, so a late pot code on an early guitar suggests a repair, and the reverse tells you the earliest possible date.

A way of working

1. Accept that a serial gives a range, and write the range down.
2. Note which readings are estimated and which are exact.
3. Add one physical clue and see whether the range shrinks.
4. If two clues disagree, say so, and do not average them.

No part of this is a judgement on authenticity. A serial that does not fit is a reason to look harder, not proof of anything.

Related

• How to read a serial number
• What a serial cannot tell you
• Timeline of formats and events