math.wiki
Number › Place value › Tally marks and Roman numerals
Open on the map
Grade 4

Tally marks and Roman numerals

Before there were digits like ours, how could people write down how many, or which year, so that anyone could read it later?

Where you are 15 in Place value

Why this exists

The oldest way to keep count is to make one mark for each thing. One sheep, one scratch. One vote, one line. This is One-to-one correspondence: each thing gets exactly one mark, so the number of marks is the number of things. These marks are called tally marks, and people still use them to count votes, laps or birds.

One mark per thing works, but a long row of marks is hard to read. Is it 17 marks or 18? So tally marks are drawn in bundles of five. Then you can count the bundles by fives and only the last few marks one at a time.

The Romans took the next step. Instead of drawing ten marks, they wrote one symbol that means ten. Instead of a hundred marks, one symbol that means a hundred. Their symbols, called Roman numerals, are still all around us. They show the year on monuments and cornerstones and at the end of movie credits. They number the chapters of some books and the hours on many clock faces and sundials. Reading them is a useful skill, and seeing how they work shows, by contrast, what is so clever about our own digits.

Definition

Tally marks. One upright mark stands for one thing. The fifth mark in each group is drawn slanting across the four before it, which closes a bundle of five. A tally is read by counting the bundles by fives and then counting the single marks by ones.

Roman numerals. Seven symbols, each with a fixed value:

SymbolValue
\mathrm{I}1
\mathrm{V}5
\mathrm{X}10
\mathrm{L}50
\mathrm{C}100
\mathrm{D}500
\mathrm{M}1{,}000

A Roman numeral is a row of these symbols, normally written from largest to smallest. Its value is found by adding the values of the symbols. For example \mathrm{CLV} = 100 + 50 + 5 = 155.

There is one exception, the subtractive pairs. When a smaller symbol is written just before a larger one, the smaller one is subtracted from the larger. Six pairs are used:

PairMeansValue
\mathrm{IV}5 - 14
\mathrm{IX}10 - 19
\mathrm{XL}50 - 1040
\mathrm{XC}100 - 1090
\mathrm{CD}500 - 100400
\mathrm{CM}1{,}000 - 100900

So \mathrm{XC} means 100 - 10 = 90, but \mathrm{CX} means 100 + 10 = 110.

The standard modern form follows these rules:

  • \mathrm{I}, \mathrm{X}, \mathrm{C} and \mathrm{M} may appear at most three times in a row.
  • \mathrm{V}, \mathrm{L} and \mathrm{D} never repeat, since two fives make a ten (\mathrm{VV} would just be \mathrm{X}), two fifties make a hundred, and two five-hundreds make a thousand.
  • Only the six pairs above subtract: \mathrm{I} before \mathrm{V} or \mathrm{X}, \mathrm{X} before \mathrm{L} or \mathrm{C}, \mathrm{C} before \mathrm{D} or \mathrm{M}.

Intuition

Here is a tally of the votes one movie got when Mandla’s class voted on what to watch. Each slanted mark closes a bundle of five.

5 10 15 20 21 22 23
Figure 1Four bundles of five and three more. Count by fives, then by ones, to get 23.

Roman numerals grew out of the same idea of bundling. Think of \mathrm{I} as a single mark. Five of them are bundled into \mathrm{V}, two fives into \mathrm{X}, five tens into \mathrm{L}, two fifties into \mathrm{C}, five hundreds into \mathrm{D}, and two five-hundreds into \mathrm{M}. The bundles go five, then two, then five, then two, over and over.

In this way the Romans could write four as \mathrm{IIII} and nine as \mathrm{VIIII}, and old Roman inscriptions often do. The short forms \mathrm{IV} (one less than five) and \mathrm{IX} (one less than ten) save symbols. The strict rules used today were settled long after Roman times, and older habits survive: many clock faces still show four as \mathrm{IIII}.

I II III IIII V VI VII VIII IX X XI XII
Figure 2Many clock faces show four as \mathrm{IIII}. The standard modern form is \mathrm{IV}.

A Roman numeral is like a pile of coins. A coin worth ten is worth ten wherever it sits in the pile. To find the total, you add up the coins. Writing the coins from largest to smallest just makes the pile easy to count, and it is what lets a reader notice the one break in the pattern: a small symbol sitting in front of a bigger one, which means “take this away”.

Results

Reading. Go from left to right. Add each symbol’s value, except that a symbol written just before a larger one is subtracted.

Writing by places. Split the number into thousands, hundreds, tens and ones, as in Expanded form, and write each part with its own pattern, largest part first. A place holding 0 writes nothing.

DigitOnesTensHundreds
1\mathrm{I}\mathrm{X}\mathrm{C}
2\mathrm{II}\mathrm{XX}\mathrm{CC}
3\mathrm{III}\mathrm{XXX}\mathrm{CCC}
4\mathrm{IV}\mathrm{XL}\mathrm{CD}
5\mathrm{V}\mathrm{L}\mathrm{D}
6\mathrm{VI}\mathrm{LX}\mathrm{DC}
7\mathrm{VII}\mathrm{LXX}\mathrm{DCC}
8\mathrm{VIII}\mathrm{LXXX}\mathrm{DCCC}
9\mathrm{IX}\mathrm{XC}\mathrm{CM}

Thousands use only \mathrm{M}: 1{,}000, 2{,}000 and 3{,}000 are \mathrm{M}, \mathrm{MM} and \mathrm{MMM}.

One standard form. Every counting number from 1 to 3{,}999 has exactly one standard Roman numeral. The largest is \mathrm{MMMCMXCIX} = 3{,}999.

No zero, no place value. Roman numerals have no symbol for zero and need none. Each symbol carries its own value, so its position never changes how much it is worth; position only decides whether it is added or subtracted.

Why it is true

Why a tally counts correctly. Each thing gets exactly one mark, so there are as many marks as things. Drawing every fifth mark across the others does not add or remove a mark; it only groups them so they can be counted by fives.

Why the place patterns work. The ones pattern uses only \mathrm{I}, \mathrm{V} and \mathrm{X}. Four and nine are one less than five and ten, so they become \mathrm{IV} and \mathrm{IX} instead of four \mathrm{I}’s in a row. The tens use \mathrm{X}, \mathrm{L}, \mathrm{C} in exactly the same way, because 10, 50, 100 are ten times 1, 5, 10. The hundreds use \mathrm{C}, \mathrm{D}, \mathrm{M}, ten times again.

Why there is only one standard form. The ones part is at most 9, the tens part at most 90, and the hundreds part at most 900. So the parts never overlap: everything after the thousands adds to less than 1{,}000, everything after the hundreds to less than 100, and so on. Reading a standard numeral, you can always tell where each part ends, and each part has only one pattern.

Why the standard form stops at 3{,}999. Thousands use only \mathrm{M}. Four thousand would need \mathrm{MMMM}, four in a row, and there is no symbol for five thousand to put after an \mathrm{M} the way \mathrm{IV} puts \mathrm{I} before \mathrm{V}. So the largest number the rules allow is \mathrm{MMM} followed by the largest hundreds, tens and ones: \mathrm{CM}, \mathrm{XC}, \mathrm{IX}.

Worked examples

The lesson is taught in two parts. Examples 1 to 3 belong to the first: tally marks, and Roman numerals that only add. Examples 4 to 7 belong to the second: the subtractive pairs, years, and why the system has no zero.

Example 1

Counting a tally

The tally in the figure above shows the votes for one movie. How many votes did it get?

Solution

There are 4 bundles and 3 single marks. Each bundle is five marks, so count the bundles by fives: 5, 10, 15, 20. Then count on by ones for the single marks: 21, 22, 23. In one line:

4 \times 5 + 3 = 20 + 3 = 23.

The movie got 23 votes. Counting by fives works because every bundle holds exactly five marks, one for each vote.

Example 2

A chapter number

Malie opens her book at chapter \mathrm{LXXVIII}. Which chapter is it?

Solution

The symbols go from largest to smallest, with no small symbol in front of a larger one, so every symbol is added:

\mathrm{LXXVIII} = 50 + 10 + 10 + 5 + 1 + 1 + 1 = 78.

It is chapter 78. Grouping the symbols helps: \mathrm{L} and \mathrm{XX} make 70, and \mathrm{VIII} makes 8.

Example 3

A year on a cornerstone

A library was built in 1886. Behrouz wants to carve the year on its cornerstone in Roman numerals. What should he carve?

Solution

Split the year into places: 1886 = 1{,}000 + 800 + 80 + 6. Write each part, largest first, using the place patterns:

  • 1{,}000 is \mathrm{M}.
  • 800 = 500 + 100 + 100 + 100 is \mathrm{DCCC}. Three \mathrm{C}’s is the most allowed.
  • 80 = 50 + 10 + 10 + 10 is \mathrm{LXXX}.
  • 6 = 5 + 1 is \mathrm{VI}.

Put the parts side by side: 1886 = \mathrm{MDCCCLXXXVI}. That is eleven symbols for a four-digit number. Check by adding: 1{,}000 + 800 + 80 + 6 = 1886.

Example 4

A number with subtractive pairs

A long movie series ends with episode \mathrm{XLIX}. Which episode is it?

Solution

Look for a smaller symbol just before a larger one. \mathrm{X} comes before \mathrm{L}, so they form the pair \mathrm{XL} = 50 - 10 = 40. \mathrm{I} comes before \mathrm{X}, so they form the pair \mathrm{IX} = 10 - 1 = 9. Add the pairs:

\mathrm{XLIX} = \mathrm{XL} + \mathrm{IX} = 40 + 9 = 49.

It is episode 49. One might try to write 49 as \mathrm{IL}, one less than fifty, but \mathrm{I} may only go before \mathrm{V} or \mathrm{X}. Writing by places avoids that slip: 49 is 4 tens and 9 ones, which is \mathrm{XL} then \mathrm{IX}.

Example 5

A year in the movie credits

At the end of an old movie, Mandla sees the year it was made: 1990. How would the credits show it in Roman numerals?

Solution

Split by places: 1990 = 1{,}000 + 900 + 90 + 0.

  • 1{,}000 is \mathrm{M}.
  • 900 is one hundred less than a thousand, the pair \mathrm{CM}.
  • 90 is ten less than a hundred, the pair \mathrm{XC}.
  • 0 ones: nothing is written.

So 1990 = \mathrm{MCMXC}.

M 1000 CM 900 + XC 90 +
Figure 3\mathrm{MCMXC} read in chunks, largest first, is 1{,}000 + 900 + 90 = 1990.

Read it back in chunks. \mathrm{M} is 1{,}000. Next \mathrm{C} stands before the larger \mathrm{M}, so \mathrm{CM} = 900. Next \mathrm{X} stands before the larger \mathrm{C}, so \mathrm{XC} = 90. The total is 1{,}000 + 900 + 90 = 1990.

Example 6

Is it standard?

Behrouz writes 99 as \mathrm{IC}, one less than a hundred. Is that the standard form? If not, what is?

Solution

It is not. \mathrm{I} may only be subtracted from \mathrm{V} or \mathrm{X}, never from \mathrm{C}. Write by places instead: 99 = 90 + 9. Ninety is \mathrm{XC} and nine is \mathrm{IX}, so

99 = \mathrm{XCIX}.

The rule keeps each subtractive pair inside one place: \mathrm{IX} lives in the ones, \mathrm{XC} in the tens. That is why every number has just one standard form.

Example 7

Where is the zero?

Write 308 and 3{,}008 in Roman numerals. Our digits need a zero in both numbers. Do the Roman numerals?

Solution

By places, 308 = 300 + 8, which is \mathrm{CCC} then \mathrm{VIII}: 308 = \mathrm{CCCVIII}. And 3{,}008 = 3{,}000 + 8, which is \mathrm{MMM} then \mathrm{VIII}: 3{,}008 = \mathrm{MMMVIII}.

Neither Roman numeral has anything for the empty places. In our number 308, the 0 holds the tens place open; without it we would write 38, a different number, because the 3 would slide into the tens place and be worth 30 instead of 300. That is Place value: a digit’s worth comes from where it sits. In \mathrm{CCCVIII} each \mathrm{C} is worth one hundred because it is a \mathrm{C}, not because of where it sits. Nothing can slide, so nothing needs holding open, and there is no zero.

The price is length and clumsy arithmetic. Our system writes every number with only ten digits, 0 to 9, used again and again in different places. The Roman system needs a new symbol for each bigger size of bundle, and adding two Roman numerals cannot be done column by column the way adding our numbers can.

Boundaries

Tally marks are for counting things one at a time. A tally can only grow by whole marks, so it shows counting numbers, not parts of a thing. Large tallies are slow to draw and to read.

The standard form covers 1 to 3{,}999. Zero has no Roman numeral, and the standard symbols cannot write 4{,}000 or more. People who needed bigger numbers used other marks, such as a bar drawn over a symbol to mean a thousand times as much, but those are outside the standard form.

Only six subtractive pairs, one small symbol each. \mathrm{IL}, \mathrm{IC}, \mathrm{VX} and \mathrm{XM} are not standard, and neither is two small symbols in front, such as \mathrm{IIX} for eight. \mathrm{V}, \mathrm{L} and \mathrm{D} are never subtracted.

Real inscriptions do not always follow the modern rules. Clock faces with \mathrm{IIII} and old carvings with \mathrm{VIIII} are still read by adding. The rules here describe the standard modern form, not every Roman numeral ever written.

No place value. A Roman symbol’s worth does not depend on its position. Place value, where the same digit is worth ten times more one place to the left, is what our system has and the Roman one does not. Place value also works with bundles of sizes other than ten, as in Numbers in other bases.

Where this goes next

Tally marks become tally charts, which record counts for several categories side by side and lead to Data tables, tallies and frequency tables and graphs. Roman numerals are worth reading for their own sake, on clocks (Telling time), monuments and chapter headings. But their main lesson is the contrast. Our digits reach any size with just ten symbols, because a digit’s place sets its worth and a zero holds an empty place. That idea carries Numbers to one million and beyond, and it is what makes rounding to a place possible: rounding needs places to round to.

Exercises

Exercises 1 to 11 go with the first part of the lesson (Examples 1 to 3), Exercises 12 to 21 with the second (Examples 4 to 7), and the reasoning group draws on both.

Part 1: tally marks and adding symbols

  1. A tally shows 3 bundles of five and 4 single marks. What number does it show?
  2. Malie tallies 32 birds. How many bundles of five and how many single marks does her tally have?
  3. A tally has 8 bundles of five and no single marks. Mandla adds 6 more marks. Describe the new tally and give its total.
  4. Write \mathrm{XXVII} in our digits.
  5. Write \mathrm{LXVI} in our digits.
  6. Write \mathrm{CCLXXXIII} in our digits.
  7. A monument shows the year \mathrm{MDCCLXXVI}. Which year is it?
  8. Write 38 in Roman numerals.
  9. Write 152 in Roman numerals.
  10. Write 2{,}615 in Roman numerals.
  11. Write 3{,}333 in Roman numerals.

Part 2: subtractive pairs and years

  1. Write \mathrm{XIV} in our digits.
  2. Write \mathrm{XCIV} in our digits.
  3. Write \mathrm{CDLX} in our digits.
  4. A cornerstone shows \mathrm{MCMLXIX}. Which year is it?
  5. A movie’s credits end with \mathrm{MMXIX}. Which year is it?
  6. Write 44 in Roman numerals.
  7. Write 279 in Roman numerals.
  8. Write the year 1904 in Roman numerals.
  9. Write 2{,}444 in Roman numerals.
  10. Write the twelve hours of a clock, 1 to 12, in standard Roman numerals. Which of them use a subtractive pair?

Reasoning

  1. Behrouz writes 45 as \mathrm{VL}. Explain why that is not the standard form, and write the standard form.
  2. A sundial shows four o’clock as \mathrm{IIII}. Explain why it can still be read, and why the standard form is \mathrm{IV}.
  3. Write 505 in Roman numerals. Explain why our number 505 needs a 0 and the Roman numeral does not, and what each 5 in 505 is worth.
  4. Write 3{,}888 in Roman numerals. Explain why no standard Roman numeral for a number from 1 to 3{,}999 has more symbols.
Answers
  1. 19. Count the bundles by fives, 5, 10, 15, then the single marks: 15 + 4 = 19.
  2. 6 bundles and 2 single marks, since 32 = 6 \times 5 + 2.
  3. 9 bundles and 1 single mark, 46 in all. The 8 bundles are 40. Five of the new marks close a ninth bundle and one is left: 40 + 6 = 46 = 9 \times 5 + 1.
  4. 27. \mathrm{XXVII} = 10 + 10 + 5 + 1 + 1 = 27.
  5. 66. \mathrm{LXVI} = 50 + 10 + 5 + 1 = 66.
  6. 283. \mathrm{CC} is 200, \mathrm{LXXX} is 80, \mathrm{III} is 3: 200 + 80 + 3 = 283.
  7. 1776. \mathrm{M} is 1{,}000, \mathrm{DCC} is 700, \mathrm{LXX} is 70, \mathrm{VI} is 6.
  8. \mathrm{XXXVIII}. 38 = 30 + 8: \mathrm{XXX} then \mathrm{VIII}.
  9. \mathrm{CLII}. 152 = 100 + 50 + 2: \mathrm{C}, \mathrm{L}, \mathrm{II}.
  10. \mathrm{MMDCXV}. 2{,}615 = 2{,}000 + 600 + 10 + 5: \mathrm{MM}, \mathrm{DC}, \mathrm{X}, \mathrm{V}.
  11. \mathrm{MMMCCCXXXIII}. Three of each of \mathrm{M}, \mathrm{C}, \mathrm{X}, \mathrm{I}, the most each may repeat.
  12. 14. \mathrm{X} = 10 and \mathrm{IV} = 4.
  13. 94. \mathrm{XC} = 90 and \mathrm{IV} = 4.
  14. 460. \mathrm{CD} = 400 and \mathrm{LX} = 60.
  15. 1969. \mathrm{M} = 1{,}000, \mathrm{CM} = 900, \mathrm{LX} = 60, \mathrm{IX} = 9.
  16. 2019. \mathrm{MM} = 2{,}000, \mathrm{X} = 10, \mathrm{IX} = 9.
  17. \mathrm{XLIV}. 44 = 40 + 4: \mathrm{XL} then \mathrm{IV}.
  18. \mathrm{CCLXXIX}. 279 = 200 + 70 + 9: \mathrm{CC}, \mathrm{LXX}, \mathrm{IX}.
  19. \mathrm{MCMIV}. 1904 = 1{,}000 + 900 + 4, and the 0 tens write nothing.
  20. \mathrm{MMCDXLIV}. 2{,}444 = 2{,}000 + 400 + 40 + 4: \mathrm{MM}, \mathrm{CD}, \mathrm{XL}, \mathrm{IV}.
  21. \mathrm{I}, \mathrm{II}, \mathrm{III}, \mathrm{IV}, \mathrm{V}, \mathrm{VI}, \mathrm{VII}, \mathrm{VIII}, \mathrm{IX}, \mathrm{X}, \mathrm{XI}, \mathrm{XII}. Only 4 and 9 use a subtractive pair: \mathrm{IV} and \mathrm{IX}.
  22. \mathrm{XLV}. \mathrm{V} is never subtracted; only \mathrm{I}, \mathrm{X} and \mathrm{C} are, and only \mathrm{X} may go before \mathrm{L}. By places, 45 = 40 + 5, which is \mathrm{XL} then \mathrm{V}.
  23. Every symbol in \mathrm{IIII} is added, 1 + 1 + 1 + 1 = 4, so it reads correctly. The standard form allows at most three \mathrm{I}’s in a row and writes four as one less than five, \mathrm{IV} = 5 - 1.
  24. 505 = \mathrm{DV}. In 505 the first 5 is worth 500 because it is in the hundreds place, and the last 5 is worth 5 because it is in the ones place; the 0 keeps the tens place open so the first 5 does not slide to the tens. In \mathrm{DV}, \mathrm{D} is always 500 and \mathrm{V} is always 5, wherever they sit, so no place needs holding.
  25. 3{,}888 = \mathrm{MMMDCCCLXXXVIII}, which is 3{,}000 + 800 + 80 + 8: \mathrm{MMM}, \mathrm{DCCC}, \mathrm{LXXX}, \mathrm{VIII}, fifteen symbols. Each place is written on its own, so the length is the sum of the lengths of the four parts. In the ones, the longest pattern is \mathrm{VIII} with 4 symbols (the others have at most 3), and the tens and hundreds patterns have the same lengths, so 8 is the longest digit there too. The thousands have at most 3 symbols, \mathrm{MMM}. So no standard numeral has more than 3 + 4 + 4 + 4 = 15 symbols, and 3{,}888 is the only one with that many.
kg1128 · updated 2026-10-20 · CC-BY-SA-4.0