Circles

Watch a bicycle wheel roll along a flat road. At every instant, the wheel touches the ground at exactly one point, and if you could freeze time and look closely, you would see that the spoke running to that touching point — the radius — always stands bolt upright, at a perfect right angle to the road. That single geometric fact, repeated at every position of a rolling wheel, is the entire engine of this chapter. A line that meets a circle at exactly one point is called a tangent, and this chapter is the formal study of tangents: what they are, how many of them exist through a given point, how their lengths behave, and how a handful of clean theorems about them unlock an entire family of board-exam problems involving circumscribed triangles and quadrilaterals. You already know a circle as the set of all points in a plane at a fixed distance (the radius) from a fixed point (the centre), with terms like chord, arc, and segment from Class IX. This chapter adds exactly one new object — the tangent — and asks three natural questions. Given a circle and a line, what are the only three things that can happen? At the point where a tangent touches a circle, how does the tangent relate to the radius drawn there (Theorem 10.1)? From a point outside a circle, how many tangents exist, and what do we know about their lengths (Theorem 10.2)? The answers are short, but their reach is long. Theorem 10.1 — the tangent is always perpendicular to the radius at the point of contact — is the single right angle that unlocks Pythagoras in almost every numerical problem here. Theorem 10.2 — the two tangents drawn from any external point to a circle are always equal in length — is the tool behind every "quadrilateral circumscribing a circle" proof you will meet. Master these two facts, and this whole chapter becomes a short, mechanical, highly scorable topic.

Key takeaways
  • A tangent to a circle is a line that touches the circle at exactly one point, called the point of contact; a line touching it at two points is a secant; a line touching it at no point is non-intersecting.
  • A tangent is the limiting case of a secant — as the two points where a secant cuts the circle slide together, the secant becomes a tangent at the instant they coincide.
  • Theorem 10.1: the tangent at any point of a circle is perpendicular to the radius through the point of contact — proved using the fact that the radius is the shortest distance from the centre to the tangent line.
  • A direct consequence: there is one and only one tangent at each point of a circle, and the radius through the point of contact is also called the normal at that point.
  • From a point, the number of tangents depends on where the point sits: zero tangents from a point inside the circle, exactly one from a point on the circle, exactly two from a point outside the circle.
  • Theorem 10.2: the lengths of the two tangents drawn from an external point to a circle are equal — proved by RHS congruence of the two right triangles formed with the centre, and equivalently by the Pythagoras theorem.
  • The line from the external point to the centre bisects the angle between the two tangents, and the angle between the two tangents is supplementary to the angle the two radii make at the centre: ∠PTQ+∠POQ=180°\angle PTQ + \angle POQ = 180°.
  • The length of a tangent from an external point PP at distance dd from the centre of a circle of radius rr is PT=d2−r2PT = \sqrt{d^2 - r^2} — a direct application of Pythagoras using Theorem 10.1's right angle.
  • Tangent-length equality forces two classic results: opposite sides of a quadrilateral circumscribing a circle add to the same total (AB+CD=AD+BCAB + CD = AD + BC), and every parallelogram circumscribing a circle is forced to be a rhombus.
  • Key idea: a tangent touches a circle at exactly one point, is always perpendicular to the radius drawn to that point, and the two tangent segments from any external point to the circle are always equal in length — two facts (Theorems 10.1 and 10.2) that solve almost every problem in this chapter.
The core insight

A tangent meets a circle at exactly one point and is always perpendicular to the radius drawn to that point (Theorem 10.1), and the two tangent segments drawn from any single external point to a circle are always equal in length (Theorem 10.2) — together these let you build a right triangle or an isosceles pair at will in any tangent problem.

What this chapter covers

The chapter is compact — just two numbered theorems — but each is load-bearing. Section 10.2, "Tangent to a Circle," sorts every possible line-circle relationship into three cases (non-intersecting, secant, tangent), defines the point of contact, and proves Theorem 10.1: the tangent at any point is perpendicular to the radius through that point. This is why "join the centre to the point of contact" is the automatic first move in almost any tangent proof. Section 10.3, "Number of Tangents from a Point on a Circle," classifies points by location (inside, on, or outside the circle) and proves Theorem 10.2: the two tangents from an external point have equal length. NCERT's own worked examples build straight on these — Example 1 uses Theorem 10.1 to show a chord of a larger circle tangent to a smaller concentric circle is bisected at the point of contact; Example 2 uses Theorem 10.2 and angle-sum reasoning to relate the angle between two tangents to the angle at the centre; Example 3 combines both with similar triangles to find an unknown tangent length. The exercises extend this two-theorem toolkit to quadrilaterals and triangles circumscribing a circle, where most of the chapter's board-exam value lives.

The core ideas

Tangent, secant, and non-intersecting line — the three possibilities

Given a circle and a line in the same plane, exactly three things can happen, and no fourth possibility exists. If they share no common point, the line is a non-intersecting line with respect to the circle. If they share exactly two common points, the line is a secant of the circle. If they share exactly one common point, the line is a tangent, and that single shared point is the point of contact. A tangent is best understood as a secant pushed to its limit: slide a secant's two intersection points toward each other along the circle, and the moment they merge into one, the secant has become a tangent. This is why NCERT calls the tangent "a special case of the secant, when the two end points of its corresponding chord coincide."

Figure
Three positions of a line relative to a circle showing a non intersecting line missing the circle entirely a secant crossing it at two points A and B and a tangent touching it at a single point A
A line and a circle — the only three possibilities non-intersecting line no common point A B secant PQ two common points A and B A tangent PQ exactly one common point A, the point of contact

Theorem 10.1 — the tangent is perpendicular to the radius at the point of contact

The rolling-wheel picture from the overview is exactly the content of the chapter's first theorem, stated formally:

Theorem 10.1: OP⊥XY,where XY is a tangent to a circle with centre O touching it at P.(1)\text{Theorem 10.1: } OP \perp XY, \qquad \text{where } XY \text{ is a tangent to a circle with centre } O \text{ touching it at } P. \qquad (1)

where: OO is the centre of the circle, PP is the point of contact, and XYXY is the tangent line at PP. The full step-by-step proof appears in the derivation block below; the underlying idea is that PP is the closest point of the line XYXY to the centre OO — every other point of the tangent line lies strictly outside the circle and is therefore strictly farther from OO than the radius OPOP. A line segment from a point to a line is shortest exactly when it is perpendicular to that line, which forces OP⊥XYOP \perp XY. Two useful remarks follow immediately: there is exactly one tangent at each point of a circle (since perpendicularity to a given radius fixes the tangent line's direction uniquely), and the radius drawn to the point of contact is sometimes called the normal to the circle at that point.

Figure
A circle with centre O and a tangent line XY touching it at point P with the radius OP drawn perpendicular to the tangent and a second point Q on the tangent joined to O showing OQ is longer than OP
Theorem 10.1 — OP is perpendicular to the tangent XY O X Y P Q Every other point Q on line XY lies outside the circle, so OQ greater than OP — OP is the shortest distance, hence OP perp XY

Number of tangents from a point, and the length of a tangent

Where a point sits relative to a circle decides how many tangents can be drawn from it. If PP lies inside the circle, every line through PP meets the circle at two points, so no tangent exists. If PP lies on the circle, Theorem 10.1's uniqueness remark gives exactly one tangent. If PP lies outside, exactly two tangents can be drawn, touching the circle at T1T_1 and T2T_2. The segment PT1PT_1 (or PT2PT_2) is called the length of the tangent from PP. NCERT's Theorem 10.2 says these two lengths are always equal:

Theorem 10.2: PT1=PT2,for any external point P and the two tangents PT1,PT2.(2)\text{Theorem 10.2: } PT_1 = PT_2, \qquad \text{for any external point } P \text{ and the two tangents } PT_1, PT_2. \qquad (2)

where: PP is any point outside the circle, and T1T_1, T2T_2 are the points of contact of the two tangents drawn from PP. Combining this with Theorem 10.1's right angle at each point of contact gives a formula used constantly in numerical problems: in right triangle OT1POT_1P, with OT1=rOT_1 = r (the radius) and OP=dOP = d (the distance from the external point to the centre), Pythagoras gives the tangent length directly,

PT=d2−r2,where d=OP>r.(3)PT = \sqrt{d^2 - r^2}, \qquad \text{where } d = OP > r. \qquad (3)

where: dd is the straight-line distance from the external point PP to the centre OO, and rr is the radius; the condition d>rd > r is exactly the statement that PP lies outside the circle, which is what guarantees a tangent (and a positive length under the square root) exists at all.

Figure
An external point P outside a circle with centre O and two tangents PT1 and PT2 drawn to the circle touching it at T1 and T2 forming a kite shape OT1PT2 with the two tangent lengths marked equal and the line OP bisecting the angle between the tangents
Theorem 10.2 — the two tangents from P are equal, PT1 = PT2 O P T1 T2 tangent PT1 tangent PT2 OP bisects angle T1PT2, and triangle OT1P is congruent to triangle OT2P (RHS), so PT1 = PT2

Angle relations and circumscribing polygons

Two further results appear in almost every exercise problem. First, since OPOP bisects the angle between the two tangents (a remark attached to Theorem 10.2's proof), the angle between the two tangents and the angle the two radii make at the centre are always supplementary:

∠T1PT2+∠T1OT2=180°.(4)\angle T_1PT_2 + \angle T_1OT_2 = 180°. \qquad (4)

where: ∠T1PT2\angle T_1PT_2 is the angle between the two tangents at the external point, and ∠T1OT2\angle T_1OT_2 is the angle subtended at the centre by the two radii to the points of contact; the full proof is given as the third derivation below. Second, when a quadrilateral ABCDABCD is drawn so that each of its four sides touches a common circle (the quadrilateral circumscribes the circle), the equal-tangent-length property applied at each of the four vertices in turn forces

AB+CD=AD+BC.(5)AB + CD = AD + BC. \qquad (5)

where: each side is broken by its point of contact into two tangent segments from the two vertices at its ends, and equal tangent lengths from each shared vertex let all eight segments be paired and regrouped into this single sum identity. This identity alone proves a parallelogram circumscribing a circle must be a rhombus: opposite sides of a parallelogram are already equal, so AB+CD=AD+BCAB+CD=AD+BC with AB=CDAB=CD and AD=BCAD=BC collapses to AB=ADAB=AD — all four sides equal.

Derivations

Derivation ATheorem 10.1 — the tangent at a point is perpendicular to the radius through that point

The goal is to prove that if XYXY is a tangent to a circle with centre OO, touching the circle at a point PP, then the radius OPOP is perpendicular to XYXY. The proof uses no trigonometry and no coordinates; it rests entirely on comparing distances from OO to different points of the line XYXY and invoking the fact that the shortest segment from a point to a line is always the perpendicular one. This is the theorem that justifies "join the centre to the point of contact" as the automatic first step of nearly every tangent problem in the chapter.

Step 1. Take any point QQ on the tangent line XYXY other than the point of contact PP, and join OQOQ. Since XYXY is a tangent, it meets the circle at PP only — every other point of the line, including QQ, must lie strictly outside the circle. (If QQ lay inside or on the circle, XYXY would cross the circle again, making it a secant, contradicting the hypothesis.)

Step 2. Because QQ lies outside the circle while PP lies on it, the distance from the centre to QQ is greater than the radius: OQ>OPOQ > OP, since OPOP equals the radius exactly while OQOQ is the distance to a point strictly farther out.

Step 3. This inequality OQ>OPOQ > OP holds for every point QQ on XYXY other than PP — Step 2 used nothing special about which point we picked. So OPOP is shorter than the segment from OO to any other point of XYXY; that is, OPOP is the shortest distance from OO to the line XYXY.

Step 4. Among all segments joining a fixed point to the various points of a given line, the shortest one is always the perpendicular segment (in any right triangle, the hypotenuse is longer than either leg, so any slanted segment is longer than the foot of the perpendicular). Since OPOP is that shortest segment (Step 3), OPOP must be the perpendicular from OO to XYXY. Hence OP⊥XYOP \perp XY, exactly the statement of Theorem 10.1.

Go deeper · Derivation B · 1 minTheorem 10.2 — the two tangents from an external point are equal in length
Derivation BTheorem 10.2 — the two tangents from an external point are equal in length

The goal is to prove that if PP is a point outside a circle with centre OO, and PQPQ, PRPR are the two tangents from PP touching the circle at QQ and RR, then PQ=PRPQ = PR. The strategy is to build two right triangles sharing a common hypotenuse and two pairs of equal sides, then invoke RHS congruence, with a Pythagoras-based shortcut given as a remark at the end for numerical work.

Step 1. Join OPOP, OQOQ, and OROR. Since PQPQ and PRPR touch the circle at QQ and RR, Theorem 10.1 applies at each point of contact: OQ⊥PQOQ \perp PQ and OR⊥PROR \perp PR, giving ∠OQP=90°\angle OQP = 90° and ∠ORP=90°\angle ORP = 90°.

Step 2. Compare triangles OQPOQP and ORPORP. Their hypotenuses coincide: OP=OPOP = OP (common to both). One pair of legs is equal because OQOQ and OROR are both radii of the same circle: OQ=OROQ = OR. Both triangles are right-angled, at QQ and RR respectively (Step 1).

Step 3. Two right triangles with equal hypotenuses and one equal pair of legs are congruent by the RHS criterion. So △OQP≅△ORP\triangle OQP \cong \triangle ORP.

Step 4. Congruent triangles have equal corresponding sides, so PQ=PRPQ = PR (CPCT). This proves Theorem 10.2. As a bonus, the corresponding angles ∠OPQ\angle OPQ and ∠OPR\angle OPR are also equal — exactly the statement that OPOP bisects the angle between the two tangents. Remark — the Pythagorean shortcut. The same result follows without congruence, directly from Theorem 10.1's right angle: in right triangle OQPOQP, PQ2=OP2−OQ2PQ^2 = OP^2 - OQ^2, and in right triangle ORPORP, PR2=OP2−OR2PR^2 = OP^2 - OR^2. Since OQ=OROQ=OR, the right-hand sides match, so PQ2=PR2PQ^2=PR^2, giving PQ=PRPQ=PR again. This route is the one to reach for whenever a problem gives you actual numbers, since it produces PT=OP2−r2PT = \sqrt{OP^2 - r^2} directly.

Go deeper · Derivation C · 1 minThe angle between two tangents is supplementary to the angle at the centre
Derivation CThe angle between two tangents is supplementary to the angle at the centre

The goal is to prove the angle relation used throughout the exercises: if PQPQ and PRPR are the two tangents from an external point PP, touching a circle with centre OO at QQ and RR, then ∠QPR+∠QOR=180°\angle QPR + \angle QOR = 180°. The proof needs nothing beyond Theorem 10.1's two right angles and the angle-sum property of a quadrilateral, applied to the four-sided figure OQPROQPR traced out by the two radii and the two tangents.

Step 1. Join OQOQ, OROR, and consider quadrilateral OQPROQPR, whose four vertices are the centre OO, the two points of contact QQ and RR, and the external point PP, taken in order. This is a genuine convex quadrilateral because QQ and RR lie on the circle while PP lies outside it, on the far side from OO.

Step 2. By Theorem 10.1, the radius to each point of contact is perpendicular to the tangent there: OQ⊥PQOQ \perp PQ gives ∠OQP=90°\angle OQP = 90°, and OR⊥PROR \perp PR gives ∠ORP=90°\angle ORP = 90° — two of the four interior angles, already known exactly.

Step 3. The angle sum of any quadrilateral is 360°360°, so ∠QOR+∠OQP+∠QPR+∠ORP=360°\angle QOR + \angle OQP + \angle QPR + \angle ORP = 360°. Substitute the two right angles: ∠QOR+90°+∠QPR+90°=360°\angle QOR + 90° + \angle QPR + 90° = 360°.

Step 4. Subtract 180°180° from both sides: ∠QOR+∠QPR=180°\angle QOR + \angle QPR = 180° — the claimed relation. This is also the engine behind NCERT's Example 2: since △QPR\triangle QPR is isosceles (PQ=PRPQ=PR by Theorem 10.2), its base angles are equal, and combining that with the supplementary relation just proved is exactly how Example 2 shows ∠QPR=2∠OQP\angle QPR = 2\angle OQP.

Worked examples

Board level - the chord of a larger circle tangent to a smaller concentric circle (extends NCERT Example 1)

Two concentric circles have radii 5 cm5\ \text{cm} and 3 cm3\ \text{cm}. Find the length of the chord of the larger circle that touches the smaller circle.

Step 1. Let OO be the common centre, and let ABAB be the chord of the larger circle (radius OA=5 cmOA = 5\ \text{cm}) that touches the smaller circle (radius 3 cm3\ \text{cm}) at the point PP. Since ABAB is tangent to the smaller circle at PP, Theorem 10.1 gives OP⊥ABOP \perp AB, with OP=3 cmOP = 3\ \text{cm}.

Step 2. Since OP⊥ABOP \perp AB and OPOP is a radius of the larger circle drawn to the chord ABAB, the perpendicular from the centre bisects the chord (a Class IX result), so AP=PBAP = PB, and AB=2 APAB = 2\,AP.

Step 3. In right triangle OPAOPA, apply Pythagoras with hypotenuse OA=5 cmOA = 5\ \text{cm} (a radius of the larger circle) and leg OP=3 cmOP = 3\ \text{cm}: AP=OA2−OP2=52−32=25−9=16=4 cmAP = \sqrt{OA^2 - OP^2} = \sqrt{5^2 - 3^2} = \sqrt{25-9} = \sqrt{16} = 4\ \text{cm}.

Step 4. Double APAP to get the full chord: AB=2×4=8 cmAB = 2 \times 4 = 8\ \text{cm}. Answer: the chord has length 8 cm\mathbf{8\ \text{cm}}.

Board level - tangent length, radius, and the Pythagorean formula (extends Exercise 10.2, Q.1)

From a point QQ, the length of the tangent to a circle is 24 cm24\ \text{cm}, and the distance of QQ from the centre is 25 cm25\ \text{cm}. Find the radius of the circle.

Step 1. Let OO be the centre and TT the point of contact of the tangent from QQ. By Theorem 10.1, OT⊥QTOT \perp QT, so △OTQ\triangle OTQ is right-angled at TT, with hypotenuse OQ=25 cmOQ = 25\ \text{cm} and one leg QT=24 cmQT = 24\ \text{cm} (the given tangent length).

Step 2. Apply Pythagoras to find the missing leg OTOT, which is the radius: OT2=OQ2−QT2=252−242=625−576=49OT^2 = OQ^2 - QT^2 = 25^2 - 24^2 = 625 - 576 = 49.

Step 3. Take the square root: OT=49=7 cmOT = \sqrt{49} = 7\ \text{cm}. Answer: the radius of the circle is 7 cm\mathbf{7\ \text{cm}}.

Board HOTS / foundation JEE-NEET - a triangle circumscribing a circle (extends Exercise 10.2, Q.12)

A triangle ABCABC is drawn to circumscribe a circle of radius 4 cm4\ \text{cm}, such that the point of contact DD on side BCBC divides it into segments BD=8 cmBD = 8\ \text{cm} and DC=6 cmDC = 6\ \text{cm}. Find the lengths of sides ABAB and ACAC.

Step 1. Let the circle touch BCBC, CACA, ABAB at DD, EE, FF. By Theorem 10.2, tangents from the same vertex are equal, so BD=BF=8 cmBD=BF=8\ \text{cm}, CD=CE=6 cmCD=CE=6\ \text{cm}, and AE=AF=x cmAE=AF=x\ \text{cm} (unknown). This gives AB=x+8AB = x+8 and AC=x+6AC = x+6.

Step 2. Set up the area two ways. With sides a=BC=14a=BC=14, b=CA=x+6b=CA=x+6, c=AB=x+8c=AB=x+8, the semi-perimeter is s=a+b+c2=x+14s = \dfrac{a+b+c}{2} = x+14.

Step 3. Split △ABC\triangle ABC into three triangles meeting at the incentre OO, each with height equal to the inradius r=4 cmr=4\ \text{cm} (since OD,OE,OF⊥OD,OE,OF \perp the sides by Theorem 10.1). So ar(△ABC)=12r(a+b+c)=12×4×2s=4s=4(x+14)\text{ar}(\triangle ABC) = \tfrac12 r(a+b+c) = \tfrac12 \times 4 \times 2s = 4s = 4(x+14).

Step 4. Equate this to Heron's formula, ar(△ABC)=s(s−a)(s−b)(s−c)\text{ar}(\triangle ABC) = \sqrt{s(s-a)(s-b)(s-c)}. With s=x+14s = x+14: s−a=(x+14)−14=xs-a = (x+14)-14 = x; s−b=(x+14)−(x+6)=8s-b = (x+14)-(x+6) = 8; s−c=(x+14)−(x+8)=6s-c = (x+14)-(x+8) = 6. So 4(x+14)=(x+14)⋅x⋅8⋅6=48x(x+14)4(x+14) = \sqrt{(x+14)\cdot x \cdot 8 \cdot 6} = \sqrt{48x(x+14)}.

Step 5. Square both sides: 16(x+14)2=48x(x+14)16(x+14)^2 = 48x(x+14). Divide through by the common factor (x+14)(x+14), valid since x+14≠0x+14 \ne 0: 16(x+14)=48x16(x+14) = 48x, i.e. 16x+224=48x16x + 224 = 48x, so 224=32x224 = 32x, giving x=7 cmx = 7\ \text{cm}. So AE=AF=7 cmAE = AF = 7\ \text{cm}.

Step 6. Substitute back into Step 1: AB=x+8=7+8=15 cmAB = x+8 = 7+8 = 15\ \text{cm} and AC=x+6=7+6=13 cmAC = x+6 = 7+6 = 13\ \text{cm}. Answer: AB=15 cmAB = \mathbf{15\ \text{cm}} and AC=13 cmAC = \mathbf{13\ \text{cm}}. Timing note: budget 6 to 7 minutes — this is the chapter's most demanding HOTS pattern, chaining Theorem 10.2's tangent-equality onto the area-via-inradius identity ar=rs\text{ar}=rs, a technique that reaches slightly beyond this chapter's own two theorems but is what examiners expect for circumscribing-triangle problems. The reusable skeleton: name the three unknown tangent segments using Theorem 10.2, express the three sides in terms of them, then bring in ar=rs\text{ar}=rs (or Heron's formula, or a right-angle condition) to pin down the last unknown.

Common traps

The most common error is skipping the first move: not joining the radius to the point of contact, so Theorem 10.1's right angle — the fact that unlocks Pythagoras — never enters the solution. A second mistake is confusing a secant with a tangent in a figure that looks like it grazes the circle; check explicitly whether one or two intersection points are implied. A third trap is writing PT=r2−d2PT=\sqrt{r^2-d^2} instead of PT=d2−r2PT=\sqrt{d^2-r^2} — dd (distance from the external point to the centre) is always the larger quantity, since PP lies outside the circle. A fourth is assuming any two tangents to a circle are equal regardless of where they are drawn from — Theorem 10.2 only guarantees equality for tangents from the same external point. A fifth, common in circumscribing-polygon problems, is forgetting to name the tangent lengths from each vertex as equal before setting up the perimeter or area equation. Finally, students mix up "angle between the tangents" (∠QPR\angle QPR) with "angle at the centre" (∠QOR\angle QOR) and treat them as equal instead of supplementary — the correct relation is ∠QPR+∠QOR=180°\angle QPR + \angle QOR = 180°.

You should be able to…

  • State precisely the difference between a tangent, a secant, and a non-intersecting line, and identify which is which from a figure.
  • State Theorem 10.1 (tangent perpendicular to the radius at the point of contact) and reproduce its shortest-distance proof.
  • Explain why there is exactly one tangent at each point of a circle, and why the radius to that point is sometimes called the normal.
  • Classify the number of tangents from a point as zero, one, or two, based on whether the point lies inside, on, or outside the circle.
  • State Theorem 10.2 (equal tangent lengths from an external point) and reproduce its RHS-congruence proof, plus the Pythagorean shortcut.
  • Compute the length of a tangent from an external point given the radius and the distance to the centre, using PT=d2−r2PT=\sqrt{d^2-r^2}.
  • Prove that the angle between two tangents from an external point is supplementary to the angle the two radii make at the centre.
  • Solve quadrilateral- and triangle-circumscribing-a-circle problems by naming equal tangent segments from each vertex first.
  • Recognise and avoid the chapter's classic traps: skipping the radius, confusing dd and rr in the Pythagorean formula, and mixing up angle equality with angle supplementarity.

Ask the tutor: Want to work through a circumscribing-quadrilateral or circumscribing-triangle proof from your own textbook, step by step, using this chapter's two theorems?

How do I actually know for certain that the tangent is perpendicular to the radius, not just approximately?

It follows from a clean distance argument, not an approximation. Take any other point Q on the tangent line besides the point of contact P; since the tangent meets the circle only at P, every such Q must lie outside the circle, so OQ (the distance from the centre) is always strictly greater than OP, the radius. That makes OP the shortest possible segment from the centre O to any point of the tangent line, and the shortest segment from a point to a line is always the perpendicular one. So the perpendicularity is an exact, provable fact — Theorem 10.1 — not a rounding or a special case. It holds at every single point of every circle, without exception.

Why are the two tangent lengths from an external point always exactly equal, not just close?

Because the two right triangles formed by the centre, the external point, and each point of contact are congruent, not merely similar. Both triangles share the same hypotenuse (the segment from the centre to the external point), have a right angle at the point of contact (Theorem 10.1), and have one leg equal to the radius (radii of the same circle are always equal). Two right triangles matching in hypotenuse and one leg are congruent by the RHS criterion, and congruent triangles have exactly equal corresponding sides — so the two tangent segments must be exactly equal, with no approximation involved. You can also see it instantly from Pythagoras: both tangent lengths equal d2−r2\sqrt{d^2-r^2} using the very same dd and rr.

When should I join the centre to a point in a tangent problem, and when should I join two external points instead?

As a rule, join the centre O to every point of contact mentioned — that activates Theorem 10.1's right angle, almost always the fastest route into Pythagoras or into congruent/similar right triangles. Join the centre to an external point (O to P or O to T) whenever the problem involves tangent length or the angle between two tangents, since that segment is the hypotenuse Theorem 10.2's proof depends on. You rarely need to join two external points directly without the centre — even comparing tangent segments along a circumscribed polygon's side is still Theorem 10.2 applied vertex by vertex.

What is the difference between the point of contact and just any point where a line crosses a circle?

A point of contact is specific to a tangent: it is the single point where a tangent line touches the circle, and by Theorem 10.1's uniqueness remark, there is exactly one tangent at that point (so the point of contact and its tangent line determine each other completely). A secant, by contrast, crosses the circle at two separate points, and neither of those two points is called a "point of contact" — that term is reserved for the one-point tangency situation. If a problem statement uses "point of contact," it is telling you a tangency is involved, which should immediately cue you to draw the radius there and mark the right angle from Theorem 10.1.

Before this

Where this leads