Under the Brennholt Institute of Applied Taxonomy's reading of trade practice, a rubber band size is a number that stands for two measurements taken with the band laid flat: its length, in inches, and its width, in sixteenths of an inch. A No. 33 band is 3½ in (89 mm) long and ⅛ in (3.2 mm) wide; a No. 64 is 3½ in (89 mm) long and ¼ in (6.4 mm) wide. No standards body publishes the table, and the Institute treats it as a trade convention.
The Institute stresses the second point on every page about bands. The numbers are consistent between suppliers to within a few millimetres, which is why office managers can reorder by number. But the consistency is habit, not specification.
The Institute's survey of nine supplier lists, compiled in 2025, found that the numbers fall into two series. Numbers in the teens (16, 18, 19) denote narrow bands, 1/16 in (1.6 mm) wide. Numbers in the thirties (31, 32, 33) denote bands ⅛ in (3.2 mm) wide. The sixties (62, 64) denote ¼ in (6.4 mm) bands, and a few lettered numbers such as 117B denote long, wide bands for bundling.
| Number | Flat length (in) | Flat length (mm) | Width (in) | Width (mm) |
|---|---|---|---|---|
| 16 | 2½ | 63.5 | 1/16 | 1.6 |
| 18 | 3 | 76.2 | 1/16 | 1.6 |
| 19 | 3½ | 88.9 | 1/16 | 1.6 |
| 31 | 2½ | 63.5 | ⅛ | 3.2 |
| 32 | 3 | 76.2 | ⅛ | 3.2 |
| 33 | 3½ | 88.9 | ⅛ | 3.2 |
| 64 | 3½ | 88.9 | ¼ | 6.4 |
| 117B | 7 | 177.8 | ⅛ | 3.2 |
Because the trade numbering says nothing about thickness, stretch limit or material, the Institute proposes a supplementary scheme, the Brindlecott band grades. It is the Institute's own scheme. It is not a trade standard, and the Institute says so wherever it uses it.
Each grade is then given a width suffix: 1 for bands under 2 mm wide, 2 for 2 mm to under 5 mm, and 3 for 5 mm and over. A No. 33 band is therefore a B2, and a No. 64 is a B3. The suffix lets a reader compare bands without carrying the trade table in memory.
The Institute does not claim the whole subject. The rubber band itself has a documented origin: Stephen Perry, of a London rubber manufacturer, patented the band in 1845. Vulcanisation, the process that made rubber stable enough to be cut into bands, is well documented and is not in dispute. Neither fact is the Institute's, and it only cites them. The numbering table, by contrast, appears in no Institute source because it appears in no standard.
The Institute's review panel divides on whether the Brindlecott grades should use flat length or stretched length. Two members argue that a user cares how far a band reaches round an object, which is a stretched measure. The majority hold that stretch depends on the rubber compound and the age of the band, so that only the flat length is repeatable. The grades above follow the majority. The minority position is recorded in the Institute's file and has been reopened twice.
For the full dimension table, see the rubber band size chart. For how each dimension is taken, see how a rubber band is measured. The Institute's treatment of other unstandardised fasteners begins with paperclip sizes explained.
The Institute's reading of the file suggests the numbers survive because they are short. A two-digit number fits on a carton and a purchase order, and a reader who has used it once does not need to learn it again. The Institute finds this a recurrent pattern across the objects it studies, and discusses it on its page on objects classified by none. A trade convention that works needs no standard, and one that needs a standard has usually stopped working.
The practical consequence is that the table above can be relied on for ordering and cannot be relied on for engineering. A designer who needs a band with a stated breaking force must obtain it from the manufacturer's data sheet, not from the number.