Learn the idea. Try it yourself.

Read the explanation or try 6 questions with hints. Free, untimed and no sign-in needed.

Read the lessonStart practising →

All lessons · Hong Kong study route

Your goal: Turn information about a square root or the first digits of a square into an exact interval. Keep the correct endpoints, find every allowed integer, and verify the final candidates without relying on rounded calculator values.

Preparation context: This original lesson supports Hong Kong preliminary short-answer preparation. It develops exact square comparisons and decimal-interval reasoning, with separately labelled proof extensions. It is a topic lesson, not an official paper, a predicted question or a complete mock exam. Every example, diagram and practice problem is newly written.

The official preliminary notice and released papers are linked at the end for exam information. A short-answer stage can reward an exact final answer, but the explanations here teach why it is correct and why no answer has been missed. The later proofs extend that reasoning; they do not change the response format of an official preliminary paper.

Before you start: You need squares of small non-negative integers, multiplication of fractions and decimals, and the meanings of < and ≤. For repair, use Exact arithmetic. Familiarity with integer bounds helps, but this lesson explains how to count the integers in an interval before introducing a compact formula.

A manageable route: Start with the three checks, then study Examples 1–2 and Problems 1–3. In a second sitting, study Examples 3–4 and Problems 4–7. Problems 8–9 are proof extensions. Allow about 25–35 minutes per sitting and another 20–30 minutes for the extensions. These are suggested study times, not contest time limits.

Three checks before using decimal prefixes

Check 1. Bracket a square root

Without finding a decimal approximation, find consecutive non-negative integers between which √70 lies.

Complete solution

Since 8² = 64 and 9² = 81, we have 64 < 70 < 81. Taking non-negative square roots gives 8 < √70 < 9. The comparison is exact. A number may be close to 8 or 9, but closeness is not needed to identify the consecutive integers.

Check 2. Square a terminating decimal exactly

Calculate 1.2² as an exact decimal or fraction.

Complete solution

Write 1.2 = 12/10. Then 1.2² = 144/100 = 1.44. Both numerator and denominator are squared. Multiplying 1.2 by 2 would give 2.4, which is doubling, not squaring.

Check 3. Read the endpoints

List the integers x satisfying 5 ≤ x < 8.

Complete solution

The permitted integers are 5, 6, 7. The symbol ≤ includes the left endpoint 5, while < excludes the right endpoint 8. There are three integers. If you included 8, mark an open circle at the right end before continuing.

If Check 1 was difficult, list the squares from 0² through 15². If Check 2 was difficult, replace a decimal by a fraction before squaring. If Check 3 was difficult, draw the endpoints and list the integers by hand. These repairs prevent most later mistakes.

Square comparisons preserve order on non-negative numbers

Throughout this lesson, √N means the non-negative square root, and N is a non-negative integer unless a problem says otherwise. When 0 ≤ a < b, the difference b² − a² = (b − a)(b + a) is positive. Thus squaring preserves the order of non-negative numbers. Conversely, comparing their squares compares the original non-negative numbers.

This is why 0 ≤ L ≤ √N < U is equivalent to L² ≤ N < U². Equality remains equality, and a strict inequality stays strict. We require non-negative bounds: squaring −4 < 3 would produce 16 < 9, which is false. Do not use the positive-number rule without its hypothesis.

Endpoint symbols carry information

For example, 5 < √N ≤ 6 becomes 25 < N ≤ 36. The value 25 is excluded and 36 is included. Do not change the symbols because you are switching from roots to squares.

An interval with 5 excluded and 6 includedAn exact number line from 4.5 to 6.5. The segment from 5 to 6 is highlighted. An open circle at 5 excludes it; a filled circle at 6 includes it.56excludedincluded5 < √N ≤ 6
Open means excluded; filled means included. Squaring the non-negative bounds gives 25 < N ≤ 36.

A decimal prefix means truncation, not rounding

When we say a non-negative number begins with 6.4, we mean that its integer part is 6 and its first digit after the decimal point is 4. Precisely, it lies in 6.4 ≤ x < 6.5. The next digit can be anything. A terminating decimal is understood to continue with zeros, so 6.4 itself has the prefix 6.4.

This is different from “rounds to 6.4 to one decimal place”. For example, 6.39 rounds to 6.4 but does not begin with 6.4, while 6.49 begins with 6.4 but rounds to 6.5. We use the usual decimal representation, with terminating decimals followed by zeros, rather than an alternative representation ending in recurring 9s.

Similarly, a prefix 2.23 means 2.23 ≤ x < 2.24. With two displayed digits after the point, the interval has width 0.01; with one, it has width 0.1. Keep any zero in the stated prefix: “8.00” is a narrower requirement than “8.0”.

Replace decimals by exact integers

Suppose the stated prefix has d digits after the decimal point. Remove the decimal point to obtain a non-negative integer a. For instance, 2.23 gives d = 2 and a = 223. Having that prefix is the condition

a/10d ≤ √N < (a + 1)/10d.

Squaring and multiplying by the positive denominator 102d gives the exact integer test

a² ≤ 102dN < (a + 1)².

The important feature is the strict upper bound. If the scaled square equals (a + 1)², the root has reached the next prefix and must be rejected. No rounded square-root display is needed.

Find the first and last permitted integer

For L ≤ N < U, first find the smallest integer at least L. Then find the largest integer strictly below U. List every integer between them. If the first is larger than the last, there are no solutions. Otherwise the number of integers is last − first + 1.

For example, 17.2 ≤ N < 20 gives 18 and 19, whereas 17 ≤ N < 20.1 gives 17, 18, 19 and 20. The +1 in the count appears because both the first and last listed integers are included.

The ceiling symbol ⌈x⌉ means the least integer at least x. Thus ⌈17.2⌉ = 18 and ⌈17⌉ = 17. In compact form, the integers in L ≤ N < U run from ⌈L⌉ to ⌈U⌉ − 1, so their count is max(0, ⌈U⌉ − ⌈L⌉). Here max(0, value) means use zero if the difference is negative. The list method is enough for every numerical problem below; the formula is a convenient summary, not a substitute for checking endpoints.

A reliable five-step routine

  1. Identify the object. Are you bounding √N, or bounding the digits of n²? Is the requested answer N, n, a list or a count?
  2. Translate the words. Write the exact interval, including total digit length when needed.
  3. Use monotonicity safely. Square non-negative bounds or compare nearby integer squares.
  4. Intersect with the other conditions. Keep only integers, and apply any parity or range restriction.
  5. Verify every boundary and completeness. Test the first and last candidates in the exact inequalities and show that the neighbouring candidates fail.

Four worked examples

Example 1. An integer part without a calculator

Find the integer part of √203.

Idea. For a non-negative real number, its integer part is the greatest integer not exceeding it. Search for consecutive squares surrounding 203.

We have 14² = 196 and 15² = 225, so 196 ≤ 203 < 225. Taking non-negative square roots gives 14 ≤ √203 < 15. Therefore the integer part is 14. There is no need to approximate √203 to several decimal places. The upper comparison proves that the answer cannot be 15, and the lower comparison proves that 14 is allowed.

Your turn 1

Find the integer part of √150 without a decimal square-root calculation.

Complete solution

Since 12² = 144 and 13² = 169, we obtain 144 ≤ 150 < 169. Hence 12 ≤ √150 < 13, and the integer part is 12. Both bounding squares are needed: checking only 12² ≤ 150 would not rule out a larger integer part.

Example 2. One decimal digit, two integer candidates

Find every non-negative integer N for which √N begins with 6.4.

Translation. The prefix gives 6.4 ≤ √N < 6.5. Since these bounds are non-negative, square them to obtain 40.96 ≤ N < 42.25.

Integer restriction. The smallest permitted integer is 41 and the largest is 42. Therefore N = 41 or 42. This is not a choice between two approximate answers: both exact integers satisfy the prefix requirement.

Exact verification. Multiplying the squared interval by 100 gives 4096 ≤ 100N < 4225. For N = 41, this is 4096 ≤ 4100 < 4225; for N = 42, it is 4096 ≤ 4200 < 4225. The neighbours 40 and 43 fail, and every other integer lies even further outside the interval. Thus the list is complete.

Your turn 2

Find every non-negative integer N for which √N begins with 3.7.

Complete solution

The prefix is equivalent to 3.7 ≤ √N < 3.8, hence 13.69 ≤ N < 14.44. The only integer is N = 14. In scaled form, 1369 ≤ 100 × 14 = 1400 < 1444 verifies it. The next lower and higher integers, 13 and 15, are outside the interval, so none has been missed.

Example 3. Two decimal digits and an exact integer test

Find every non-negative integer N for which √N begins with 2.23.

Bridge from Example 2. Two decimal places make the prefix interval ten times narrower. We need 2.23 ≤ √N < 2.24, not an interval based on rounding to two decimal places.

Write the bounds as 223/100 and 224/100. Squaring gives 49729/10000 ≤ N < 50176/10000, or 4.9729 ≤ N < 5.0176. The only candidate is N = 5. Check it without decimal approximation: 49729 ≤ 10000 × 5 = 50000 < 50176. This verifies that √5 has the required prefix.

The interval is shorter than 1, so it can contain at most one integer, but it need not contain any. An interval’s small width alone does not prove existence. The exact candidate test is the missing step.

Your turn 3. An empty interval

Is there a non-negative integer N for which √N begins with 5.01? Find all such N.

Complete solution

The prefix means 5.01 ≤ √N < 5.02. Squaring gives 25.1001 ≤ N < 25.2004. This entire interval lies strictly between the consecutive integers 25 and 26, so there is no such integer N. Rounding an endpoint to 25 would admit a false candidate: √25 = 5.000…, which does not begin with 5.01.

Leading digits of an integer need a total length

The positive integer 364 and the positive integer 36400 both begin with 36, but have different sizes. If an integer has exactly five digits and begins with 36, it lies from 36000 through 36999 inclusive. It is convenient to write the same condition as 36000 ≤ M < 37000.

More generally, if M has exactly D digits and its first r digits form the positive r-digit integer p, then p × 10D−r ≤ M < (p + 1) × 10D−r. We assume 1 ≤ r ≤ D and no leading zero in p. Appending D − r zeros gives the smallest number with that prefix; increasing the prefix by one reaches the first number outside it.

Example 4. The square is bounded; the answer is its root

Find every positive integer n such that n² has exactly five digits and begins with 36.

Translate the digit condition. We need 36000 ≤ n² < 37000. Since n is positive, squares increase as n increases. Compare nearby squares at both ends.

At the lower end, 189² = 35721 < 36000, while 190² = 36100 ≥ 36000. At the upper end, 192² = 36864 < 37000, while 193² = 37249 ≥ 37000. Therefore n = 190, 191 or 192. The middle square is 191² = 36481, so it also has five digits and the required prefix.

Completeness. Every smaller positive integer has square below 36000, and every larger one has square at least 37000. The question asks for n, not n²; giving only 36100, 36481 and 36864 would answer a different question. Negative roots are excluded by the stated domain.

Your turn 4

Find every positive integer n such that n² has exactly four digits and begins with 12.

Complete solution

Four digits with prefix 12 means 1200 ≤ n² < 1300. The neighbouring squares are 34² = 1156, 35² = 1225, 36² = 1296 and 37² = 1369. Thus n = 35 or 36. Squares increase for positive integers, so the failed neighbours 34 and 37 rule out every smaller or larger positive n. Both surviving squares have the specified length and prefix.

Common mistakes and a repair

  • Using rounding instead of a prefix. Write the interval first. The prefix 6.4 means [6.4, 6.5), regardless of how many later digits the number has. Square brackets mean include an endpoint; a round bracket means exclude it.
  • Rounding a squared endpoint. Keep terminating decimals exact, or multiply by 102d. A rounded endpoint can incorrectly add or remove an integer.
  • Including the next prefix. The upper inequality is strict. A root equal to 2.24 no longer begins with 2.23.
  • Ignoring an included lower endpoint. The prefix 8.00 includes √64 = 8.000…. A perfect square can be a valid endpoint.
  • Forgetting digit length. “Begins with 36” alone does not imply 36000 ≤ M < 37000. Five digits is the additional information that fixes the scale.
  • Squaring across negative bounds without care. First check non-negativity. The order argument in this lesson applies on the non-negative side.
  • Proving at most one and claiming exactly one. An interval of width less than 1 may be empty. Exhibit a candidate and verify it.
  • Stopping after one candidate. Find the first and last allowed integers, then account for all integers between them. Apply parity or other restrictions afterwards.

Nine original written problems

Problems 1–3 build fluency; 4–7 change an endpoint, a prefix or an additional condition; 8–9 ask for short proofs. Open Hint 1 before Hint 2, then write your own complete solution before reading ours. All decimal prefixes below have the interval meaning defined above.

Foundation 1. The integer part

Find the integer part of √130 without evaluating a decimal square root.

Hint 1

Find consecutive squares around 130.

Hint 2

Compare 11² and 12² with 130.

Complete solution

We have 11² = 121 and 12² = 144, so 121 ≤ 130 < 144. Since square roots are non-negative, 11 ≤ √130 < 12. The greatest integer not exceeding √130 is therefore 11. The upper bound rules out 12 and every larger integer.

Foundation 2. A one-digit prefix

Find every non-negative integer N for which √N begins with 7.2.

Hint 1

Write the interval from 7.2 up to, but not including, 7.3.

Hint 2

Square both endpoints exactly: 7.2² = 51.84 and 7.3² = 53.29.

Complete solution

The requirement is 7.2 ≤ √N < 7.3, equivalent to 51.84 ≤ N < 53.29. Hence N = 52 or 53. In integer form, the test is 5184 ≤ 100N < 5329; it holds for 5200 and 5300. The integers 51 and 54 fail on opposite sides, so the two listed values are all the possibilities.

Foundation 3. Preserve two decimal places

Find every non-negative integer N for which √N begins with 1.41.

Hint 1

Use the exact fractions 141/100 and 142/100.

Hint 2

After squaring, multiply the entire inequality by 10000.

Complete solution

We need 1.41 ≤ √N < 1.42. This gives 19881 ≤ 10000N < 20164, or 1.9881 ≤ N < 2.0164. The only candidate is N=2N = 2, and 19881 ≤ 20000 < 20164 verifies it. No other integer lies between the squared bounds. A rounded value of √2 is unnecessary.

Transfer 4. The included perfect square

Find every non-negative integer N for which √N begins with 8.00.

Hint 1

A terminating decimal continues with zeros in our definition.

Hint 2

The interval starts at 8.00, with that endpoint included.

Complete solution

The prefix means 8.00 ≤ √N < 8.01. Squaring yields 64 ≤ N < 64.1601. Thus N = 64 is the only integer. It is valid because √64 = 8.000…, so the lower endpoint really has the required prefix. The next integer 65 is already above the upper squared bound.

Transfer 5. Reverse the endpoint choices

How many non-negative integers N satisfy 5 < √N ≤ 6?

Hint 1

Square both bounds, preserving < and ≤ exactly.

Hint 2

The first allowed integer is one more than 25; the last is 36.

Complete solution

Squaring the non-negative bounds gives 25 < N ≤ 36. The integers are 26, 27, …, 36, so there are 36 − 26 + 1 = 11. The value 25 must be excluded because its root equals 5, while 36 must be included because its root equals 6. Counting the endpoints as if the interval were 25 ≤ N < 36 would give the same count accidentally but the wrong set; the actual inequalities matter.

Transfer 6. A missing square prefix

Find every positive integer n such that n² has exactly four digits and begins with 45.

Hint 1

Translate the digit condition to 4500 ≤ n² < 4600.

Hint 2

Compare the consecutive squares 67² and 68².

Complete solution

The required interval is 4500 ≤ n² < 4600. But 67² = 4489 lies below it, and 68² = 4624 lies above it. For positive integers, squares strictly increase; there is no integer strictly between 67 and 68. Therefore there is no such positive integer n. The interval contains many integers, but none is a perfect square. This differs from an empty interval for N in a square-root-prefix problem.

Transfer 7. A prefix and parity

Find every positive even integer n such that n² has exactly five digits and begins with 20.

Hint 1

First find all positive integer roots whose squares lie in the prefix interval.

Hint 2

Use 141², 142², 144² and 145² to identify the endpoint candidates; then apply evenness.

Complete solution

The digit condition is 20000 ≤ n² < 21000. We have 141² = 19881, 142² = 20164, 144² = 20736 and 145² = 21025. Thus the allowed positive integers before imposing parity are 142, 143 and 144; the middle square 143² = 20449 also lies inside the interval. Keeping only even values gives n = 142 or 144. The failed neighbouring squares prove completeness, and the parity condition removes 143 without changing the endpoint reasoning.

Extension 8. When does a zero digit force a square?

For which positive integers k is N = k² the only non-negative integer N whose square root has integer part k and first decimal digit 0? Prove that your list of k is complete.

Hint 1

Translate the condition to k ≤ √N < k + 0.1.

Hint 2

The squared interval is k² ≤ N < k² + (2k/10) + 1/100. Ask when k² + 1 first fits.

Complete solution

The prefix condition is k ≤ √N < k + 1/10. To square the upper bound, multiply the two brackets: (k + 1/10)² = (k + 1/10)(k + 1/10) = k² + k/10 + k/10 + 1/100 = k² + 2k/10 + 1/100 = k² + k/5 + 1/100. Thus k² ≤ N < k² + k/5 + 1/100. The value N = k² always works. For 1 ≤ k ≤ 4, the extra width is at most 4/5 + 1/100 = 81/100 < 1, so no integer larger than k² lies in the interval. Conversely, if k ≥ 5, the width is at least 1 + 1/100 > 1. Then N = k² + 1 also lies strictly below the upper bound and has the required prefix. Therefore the complete answer is k = 1, 2, 3, 4. The proof handles every larger k together rather than testing a finite sample.

Extension 9. A sharp guarantee of existence

Prove that every one-decimal-place prefix at least 5.0 occurs as the prefix of √N for some non-negative integer N. Show also that the prefix 4.9 does not occur. A one-decimal-place prefix here is a/10 for an integer a, with its one decimal digit retained.

Hint 1

Write the squared interval using the integer a. Its length is (2a + 1)/100.

Hint 2

Any interval L ≤ N < U of length greater than 1 contains an integer: take the least integer at least L.

Complete solution

For a ≥ 50, the prefix interval gives a²/100 ≤ N < (a + 1)²/100. Its length is ((a + 1)² − a²)/100 = (2a + 1)/100 ≥ 101/100 > 1. Set L = a²/100 and U = (a + 1)²/100, and choose the least integer N ≥ L. If N ≥ L + 1, then the smaller integer N − 1 would still satisfy N − 1 ≥ L, contradicting the choice of N as the least such integer. Hence L ≤ N < L + 1. Since U − L > 1, we have L + 1 < U, so L ≤ N < L + 1 < U. This proves that N satisfies both required inequalities. Hence a suitable N exists for every a ≥ 50. For the prefix 4.9, however, the squared interval is 24.01 ≤ N < 25. The integer 24 is too small and 25 is excluded, so there is no integer N. Thus all the stated prefixes from 5.0 onward occur, while the preceding one 4.9 does not. This proves the threshold for the stated guarantee, not that every smaller prefix fails.

A proof and solution rubric

This checklist is for self-review, not an official marking scheme or automatic proof grading.

  1. Translation: State the precise object, domain and interval. Preserve prefix length, total digit length and endpoint symbols.
  2. Justification: Explain why squaring or comparing squares preserves the inequalities. Keep computations exact.
  3. Completeness: Identify the first and last candidates, or give an argument covering every case. For an existence claim, exhibit a candidate or prove an integer lies in the interval.
  4. Verification: Recheck boundary candidates and all extra restrictions. State whether the answer is a root, a radicand, a list or a count.

A correct numerical guess is a starting point. To improve it, add the inequality that admits the candidate and the bounds that exclude all others. In a proof, a table of several examples supports a conjecture but does not establish a claim about every positive integer.

Practise exact decisions

The adaptive checks below move among easier and harder questions on bounds, prefixes and integer candidates. If feedback identifies a missed endpoint or a rounding error, follow the explanation link before trying again. Written proofs in Problems 8–9 still require your own reasoning; they are not automatically graded.

A completed session means six responses, including any requested solutions. It does not mean six independently correct answers or mastery. Return later to a solution-assisted question without opening its explanation. Practice remains free and uses the site’s existing privacy and local-progress choices; no written proof or personal details are required.

Interactive practice loads here. You can also use the complete question set below.

Explain the boundaries before you leave

Without reopening the theory, explain why the next prefix is excluded, why squaring requires non-negative bounds, and why an interval narrower than 1 need not contain an integer. Redo one problem that required a hint and check it using an exact scaled inequality. Note privately whether your new attempt was independent or solution-assisted.

Return to Hong Kong’s study and paper route · Repair exact arithmetic · Choose another lesson

Preparation sources and original material

Use HKAGE’s official 2026 preliminary notice and IMO Hong Kong Committee’s released tests for official contest information and papers. These links provide preparation context; all examples and practice questions here are original, not reproduced or translated past-paper questions.

All teaching text, the interval diagram and the problems on this page are original IMOolympiad.com learning material. This independent educational website is not an official organiser and does not claim affiliation or endorsement. Preparation context checked 28 October 2026.