A measured number carries a hidden promise
Suppose you weigh a coin and your balance reads 5.0 grams. Now suppose a better balance reads 5.00 grams. Mathematically 5.0 and 5.00 are the same value — but to a chemist they say two different things. The first promises "I know this to a tenth of a gram." The second promises "I know this to a hundredth." Those trailing zeros are not decoration; they are a claim about how sure you are. The digits in a number that actually carry that information are its significant figures.
So the digits you write down are an honesty contract. They tell the next person how far they may trust your result. Writing one digit too few throws away hard-won information; writing one too many is a quiet lie — pretending you measured something you never did. This is why significant figures sit right at the heart of every quantitative analysis: they are how the field keeps its results truthful about their own uncertainty.
Counting significant figures: the zero puzzle
Any non-zero digit always counts — 1 through 9 are always significant. The trouble is only ever the zeros, because a zero can do two different jobs. Sometimes a zero genuinely reports a measurement (it is a digit you really know). Other times a zero is just a placeholder, holding the decimal point in the right spot. The skill is telling those two kinds of zero apart.
- Zeros *between* non-zero digits always count. In 4.05, that middle zero is significant — three sig figs.
- Leading zeros (at the front, before any non-zero digit) never count. In 0.0042, the zeros are just placeholders — two sig figs (the 4 and the 2).
- Trailing zeros *after* a decimal point always count. In 3.200, both final zeros are real measured digits — four sig figs.
- Trailing zeros in a whole number with no decimal point are ambiguous: '1500' could be two, three, or four sig figs. Scientific notation removes the doubt.
That last case is exactly where scientific notation earns its keep. Writing 1500 leaves a reader guessing how many zeros you really measured. But 1.5 × 10³ says clearly "two significant figures," while 1.500 × 10³ says "four." The exponent carries the size, and the part in front carries — unambiguously — your honest digits. Many chemists default to scientific notation for precisely this reason: it cannot lie about how much you know.
Rounding without lying
When a calculation spits out 3.847291 but you only honestly know three figures, you must trim it — that trimming is rounding. The everyday rule is simple: look at the first digit you are dropping. If it is 5 or more, round the kept digit up; if it is 4 or less, leave it. So 3.847291 to three sig figs becomes 3.85 (the dropped 7 pushes the 4 up to 5). Rounding is not sloppiness — done right, it is the act of stating your result at exactly the honesty level your measurement supports.
Two cautions save beginners real grief. First, round only at the very end of a calculation — keep extra digits while you work, and trim once at the finish. Rounding partway through, then feeding the rounded number into the next step, lets small errors snowball. Second, when several numbers are added or multiplied, your answer can be no more precise than the least precise number that went into it. A chain is only as strong as its weakest link, and a result is only as trustworthy as its shakiest measurement.
Significant figures vs precision: a gentle distinction
It is worth heading off a common mix-up. Significant figures are about *how you write a number down* — a notation, a way of communicating on paper. Precision (which the next guides will treat fully) is about *how repeatable the actual measurement is* — a property of your instrument and technique. They are tightly related: an honest count of significant figures should reflect the real precision of the determination. But they are not the same thing. Writing more digits does not make a measurement more precise; it only claims more — and an empty claim is exactly what good practice forbids.