The International System of Units (SI) is the modern metric system and the near-universal standard for reporting quantitative data in scientific manuscripts. Most researchers know the SI loosely from school — metres, kilograms, seconds — but manuscript preparation exposes a second layer most people never learn: a detailed set of writing rules governing spacing, capitalization, pluralization, and symbol formatting that journals enforce and that authors routinely get wrong. This guide covers both layers: what the SI actually is, including the 2019 redefinition of its base units, and the specific formatting conventions that matter when you are writing up results for submission.
What the SI is, and who maintains it
The International System of Units is maintained by the International Bureau of Weights and Measures (BIPM), based in Sèvres, France, under the authority of the General Conference on Weights and Measures (CGPM — Conférence générale des poids et mesures), the intergovernmental body established by the 1875 Metre Convention. The CGPM meets periodically to approve revisions to the SI; national metrology institutes such as NIST (United States) and NPL (United Kingdom) then publish their own guidance translating BIPM’s definitions into practical use.
The authoritative reference document is the SI Brochure, 9th edition (2019), published by BIPM. It is the single source that defines the base units, the rules for forming derived units, and the typographic conventions for writing unit symbols and names. When a journal’s style guide is silent or ambiguous on a units question, the SI Brochure is the document to consult.
Quantity, unit, and dimension: three distinct concepts
Manuscript reviewers occasionally flag confusion between three related but distinct ideas:
- A quantity is a property that can be measured — length, mass, time, temperature, electric current, amount of substance, luminous intensity.
- A unit is a defined reference value used to express the magnitude of a quantity — the metre is a unit of the quantity length.
- A dimension describes how a quantity relates to the base quantities algebraically (for example, speed has the dimension length/time), independent of which units are chosen to express it.
Quantity symbols (like m for mass or t for time) are set in italic type; unit symbols (like m for metre or s for second) are set in roman (upright) type. This distinction is a frequent copyediting flag — see the formatting rules below.
The seven base units and the 2019 redefinition
The SI is built from seven base units, each corresponding to one base quantity:
| Base quantity | Unit name | Symbol |
|---|---|---|
| Time | second | s |
| Length | metre (meter) | m |
| Mass | kilogram | kg |
| Electric current | ampere | A |
| Thermodynamic temperature | kelvin | K |
| Amount of substance | mole | mol |
| Luminous intensity | candela | cd |
Since 20 May 2019, all seven base units have been defined by fixing the numerical values of seven defining constants, rather than by physical artefacts or laboratory-specific procedures. This was the most significant revision to the SI since its founding. The seven defining constants are:
- The caesium hyperfine transition frequency, ΔνCs (defines the second)
- The speed of light in vacuum, c (defines the metre, in combination with the second)
- The Planck constant, h (defines the kilogram)
- The elementary charge, e (defines the ampere)
- The Boltzmann constant, k (defines the kelvin)
- The Avogadro constant, NA (defines the mole)
- The luminous efficacy of a defined monochromatic radiation, Kcd (defines the candela)
The most publicized consequence of this redefinition was the retirement of the international prototype of the kilogram (IPK) — the physical platinum-iridium cylinder held at BIPM that had served as the literal definition of the kilogram since 1889. Because the IPK could gain or lose microscopic mass over time through handling and cleaning, tying the kilogram’s definition to it meant the unit itself was not perfectly stable. Defining the kilogram instead through a fixed value of the Planck constant (realized in practice using instruments such as the Kibble balance) removed that dependence on a single physical object. For the practising researcher, none of this changes how a kilogram is used day to day — the point of the redefinition was that the everyday value of each base unit did not change at the moment of redefinition, only the method by which it is defined and realized.
Derived units
Every other SI unit is a derived unit, formed by multiplying or dividing base units. Some derived units carry special names and symbols, generally honouring a scientist, because writing out the base-unit combination every time would be unwieldy. Common examples relevant to research writing:
| Quantity | Unit name | Symbol | Expressed in base/other SI units |
|---|---|---|---|
| Frequency | hertz | Hz | s−1 |
| Force | newton | N | kg·m·s−2 |
| Pressure, stress | pascal | Pa | N/m2 |
| Energy, work, heat | joule | J | N·m |
| Power, radiant flux | watt | W | J/s |
| Electric charge | coulomb | C | A·s |
| Electric potential | volt | V | W/A |
| Celsius temperature | degree Celsius | °C | K (offset scale) |
| Catalytic activity | katal | kat | mol·s−1 |
| Radioactivity (activity of a radionuclide) | becquerel | Bq | s−1 |
| Absorbed dose (ionizing radiation) | gray | Gy | J/kg |
| Equivalent dose (ionizing radiation) | sievert | Sv | J/kg |
Note that the gray and the sievert have the same base-unit expression (J/kg) but measure different quantities — absorbed dose versus biologically weighted equivalent dose — and using one where the other is meant is a real and consequential error in clinical and radiation-safety writing, not just a stylistic slip.
SI prefixes, including the 2022 additions
Prefixes scale a unit up or down by powers of ten without changing the underlying quantity. The core prefix set runs from yocto (10−24) to yotta (1024). In November 2022, the 27th CGPM approved four new prefixes — the first expansion to the SI prefix range since 1991 — extending the range to accommodate very large data-storage and cosmological quantities and very small quantities at the other extreme:
| Prefix | Symbol | Factor |
|---|---|---|
| quetta | Q | 1030 |
| ronna | R | 1027 |
| yotta | Y | 1024 |
| zetta | Z | 1021 |
| exa | E | 1018 |
| peta | P | 1015 |
| tera | T | 1012 |
| giga | G | 109 |
| mega | M | 106 |
| kilo | k | 103 |
| milli | m | 10−3 |
| micro | µ | 10−6 |
| nano | n | 10−9 |
| pico | p | 10−12 |
| femto | f | 10−15 |
| atto | a | 10−18 |
| zepto | z | 10−21 |
| yocto | y | 10−24 |
| ronto | r | 10−27 |
| quecto | q | 10−30 |
Two rules the SI Brochure is explicit about: prefixes are never stacked (write nm, never mµm), and a prefix symbol is joined directly to the unit symbol with no space (5 km, not 5 k m). For mass specifically, the kilogram is the base unit despite carrying a prefix already, so smaller masses are prefixed onto the gram (milligram, mg) rather than the kilogram.
Writing rules manuscripts get wrong
This is the section with the most day-to-day relevance for manuscript preparation. Reviewers, copyeditors, and production staff at SI-compliant journals check for these specifically.
- Space between the number and the unit symbol. Write 25 mg, not 25mg. The SI Brochure specifies a non-breaking space (to prevent a line break stranding the number from its unit). The two exceptions are the degree, minute, and second of plane angle (°, ′, ″) and the degree Celsius, which are written closed up to the number: 37°C, not 37 °C.
- Unit symbols are never pluralized. Write 10 kg, not 10 kgs. Unit names, when spelled out in running text, do take a normal plural: “ten kilograms” is correct, “ten kilogram” is not.
- Unit symbols take no full stop, except where normal sentence punctuation requires one (i.e., a period ending a sentence). “5 mL.” mid-sentence is wrong; “the sample volume was 5 mL.” at the end of a sentence is fine because that period belongs to the sentence, not the unit.
- Roman (upright) type for unit symbols, italic for quantity symbols. The unit symbol “m” (metre) is roman; the quantity symbol “m” (mass) is italic. Word-processing autoformat and LaTeX math mode both routinely italicize unit symbols by accident when they appear inside an equation — a common source of inconsistency that copyeditors flag.
- Capitalization: unit names are lowercase even when derived from a person’s name; the corresponding symbol is capitalized. Write “newton” (not “Newton”) for the unit name, but “N” (capital) for its symbol; “kelvin” (not “Kelvin”), symbol “K.” This trips up authors because it runs against the instinct to capitalize a name.
- Litre: “L” versus “l.” Both the lowercase “l” and uppercase “L” are accepted symbols for the litre under the SI Brochure, but because lowercase “l” is easily confused with the numeral “1” in many typefaces, NIST and most journals in practice prefer the capital “L” (5 L, 5 mL) even though this is an exception to the usual rule that symbols not derived from a proper name are lowercase.
- The solidus (/) and negative exponents. A compound unit can be written with a slash (m/s) or with a negative exponent (m·s−1), but the SI Brochure specifies that a given expression should never contain more than one solidus — write kg·m−1·s−2 or kg/(m·s2), never kg/m/s2, which is ambiguous about operator precedence.
- Decimal marker and digit grouping. The SI Brochure permits either a comma or a point as the decimal marker, reflecting genuine international variation, but is explicit that whichever is used, digits should be grouped in threes using a thin space (not a comma or point) to avoid the exact ambiguity that mixing conventions creates: 15 739.012 03, not 15,739.01203. Most English-language journals default to a decimal point and a thin space (or no separator at all for four-digit numbers) rather than the comma-grouping common in general English prose.
Non-SI units accepted for use, and units you’ll still see in older or field-specific literature
The SI Brochure explicitly accepts a short list of non-SI units for continued use alongside SI units because of their practical importance or historical entrenchment: the minute, hour, and day (time); the degree, minute, and second (plane angle); the litre (volume); the tonne (mass); the electronvolt (energy, common in particle and atomic physics); the dalton or unified atomic mass unit (mass, common in chemistry and molecular biology); and the astronomical unit (distance, in astronomy). These are accepted precisely because forcing every field to write everything exclusively in coherent SI units (joules instead of electronvolts, for example) would make the literature harder to read without a compensating gain in rigor.
Older literature, and some ongoing field conventions, also use units the SI Brochure does not endorse for general use, including the ångström (still common in crystallography for interatomic distances), the calorie (nutrition science and some older biochemistry), the bar (some atmospheric and engineering contexts, close to but not identical to the pascal-derived unit), and units like the erg, dyne, or gauss from the older CGS system. Authors working from historical sources or replicating an older study’s methods section will encounter these and should convert to SI equivalents, or provide both, per the target journal’s policy.
Journal and discipline variation
General SI usage is the default expectation, but discipline convention and journal house style both override the general rule in specific, well-established cases:
- Clinical medicine commonly reports blood pressure in millimetres of mercury (mmHg) rather than the SI-coherent pascal, and many clinical laboratories still report certain analytes (glucose, cholesterol) in mg/dL rather than the SI-preferred mmol/L, particularly in North American practice, even though international and most non-US journals expect SI units for lab values.
- Nutrition science commonly retains the calorie (or kilocalorie, informally “Calorie”) for energy content rather than the joule, reflecting both regulatory labelling requirements and long-standing reader familiarity.
- Crystallography and structural biology commonly retain the ångström (0.1 nm) for bond lengths and unit-cell dimensions rather than the nanometre or picometre.
- Astronomy retains the astronomical unit, light-year, and parsec for distance rather than expressing everything in metres, where the numbers involved would be unwieldy.
None of this is a license to use whatever unit an author prefers: it reflects specific, named exceptions that a field’s leading journals have converged on. The decisive source for any given submission is the target journal’s own author guidelines — most SI-compliant journals state explicitly whether SI units are mandatory, when a non-SI unit may be used alongside an SI value in parentheses, and how they want compound units formatted. When journal guidance and general SI convention conflict, the journal’s instructions to authors govern for that submission. Discipline-specific style manuals also address units directly — for example, chemistry manuscripts following ACS style conventions and social-science manuscripts following APA style both specify how numbers and units should be formatted in running text and tables.
Significant figures and uncertainty
A unit only reports a magnitude; it says nothing about how precisely that magnitude is known. Reporting a measured quantity to more significant figures than the measurement justifies implies a false precision, and reporting to fewer discards real information. The same discipline applies to units and values shown in tables and figures — see CASRAI’s guide on designing effective scientific figures for how axis labels, units, and error bars should be presented consistently with the units conventions covered here. As a general rule, the number of significant figures in a reported value should reflect the resolution and uncertainty of the measurement or instrument that produced it, and a stated uncertainty (standard deviation, standard error, or confidence interval, as appropriate to the analysis) should accompany reported quantities where precision matters to interpretation — a mean without an associated measure of variability or precision is, on its own, an incomplete result in most quantitative reporting contexts.
Frequently asked questions
What does SI stand for?
SI stands for Système international d’unités (International System of Units), the name given to the modern metric system by the CGPM in 1960.
Is the metric system the same as SI?
SI is the internationally standardized, modern form of the metric system. “Metric system” is sometimes used loosely to include older metric variants (like the CGS system) that predate or fall outside current SI conventions, so the two terms are closely related but not strictly interchangeable.
Why was the kilogram redefined in 2019?
The kilogram was, until 2019, the only base unit still defined by a physical artefact — a platinum-iridium cylinder held at BIPM. Because that artefact’s mass could drift very slightly over time through handling, cleaning, and environmental exposure, the definition itself was not perfectly stable. Redefining the kilogram in terms of a fixed value of the Planck constant removed that dependence on a single physical object.
Do unit symbols ever take a plural “s”?
No. Unit symbols are invariant regardless of the numerical value: 1 kg and 10 kg, never 10 kgs. Only spelled-out unit names take an ordinary plural in running text.
Should I use “L” or “l” for litre?
Both are valid SI symbols, but most journals and NIST guidance prefer the capital “L” because a lowercase “l” is easily misread as the numeral “1” in many fonts.
Can I use non-SI units like mmHg or calories in a manuscript?
Some non-SI units are in wide, accepted use in specific fields (mmHg in clinical blood pressure reporting, calories in nutrition) even though the SI Brochure does not include them among its explicitly accepted exceptions. Whether a given journal allows this, requires an SI equivalent in parentheses, or requires SI units exclusively depends on that journal’s author guidelines, which should always be checked before submission.
