Chapter II · Functions of real variables
Polar coordinates
Locating a point by its distance and direction from the origin, polar equations of lines, circles and conics, and why Hardy insists that is never negative.
Why this matters
Some curves are awkward in and but simple when described by distance and direction from a point: circles through the origin, spirals, conics about a focus. §22 introduces these polar coordinates and the polar equation of a locus. It also contains, in a footnote, a convention that matters more than it looks: for Hardy the distance is never negative. The same equation can draw half a curve or the whole of it depending on that convention.
The idea
Instead of walking along one axis and along the other, stand at the origin , turn through an angle from the positive axis of , and walk a distance . The pair fixes the point , and right-angled trigonometry links the two descriptions:
A polar equation such as should be read as an instruction: for each direction , the points of the locus in that direction are at the distances that the equation allows. Sweep a ray round and mark where the equation puts points on it.
Try this. With , drag slowly from to and watch where the point on the ray is. Pay attention to what happens between and . Then switch on “allow negative ” and sweep again.
- points with r ≥ 0
- the ray at angle θ
What to notice
For the equation gives a positive and the point runs once round a circle through . For it gives , which in Hardy’s convention is no point at all: the ray misses the curve. If negative is allowed (plotted backwards along the ray), those directions retrace the same circle a second time.
Precise statements
Let be a point, its distance from the origin, and the angle from the positive axis of to , measured anticlockwise, with . Then and are the polar coordinates of , and
A relation between and , such as , satisfied by exactly the points of a locus, is its polar equation.
At the origin itself and is not determined; any will do. In Hardy’s convention is essentially positive: a polar equation that gives a negative value of at some gives no point in that direction.
(i) Let be the foot of the perpendicular from to a line not through , with and in direction . The polar equation of the line is
(ii) The circle with radius whose centre has polar coordinates has polar equation
In particular (, ), is a circle of radius through the origin.
Proofs
A point is on the line exactly when its projection onto the direction of has length . Express that projection in polar coordinates.
The line in cartesian form
The unit vector in direction is , and the line is perpendicular to it at distance from . So is on the line exactly when the component of along equals : that is, .
Substitute ,
, by the subtraction formula for the cosine.
Conclude
So is on the line exactly when . Since and , only directions with , that is , meet the line, as they should: those are the directions that point towards it.
Apply the cosine rule to the triangle formed by , the centre and the point .
The triangle
, , and the angle between and is (its cosine is all that matters, so the sign of the difference is irrelevant).
Cosine rule
. (If , , are collinear the formula still holds, as you can check directly.)
Conclude
is on the circle exactly when , that is, . With and this becomes , and dividing by (allowed for ) gives . The origin itself lies on this circle: it is the point with , given by .
The general circle has a surprise. Solving the quadratic for gives
so when is outside the circle a ray can meet it twice: is then a two-valued function of , defined only for the directions that hit the circle.
Try this. With the general circle, keep the origin outside it () and sweep , counting the points on the ray; then make and sweep again. Then choose the conic with , sweep , and finally allow negative .
- points with r ≥ 0
- the ray at angle θ
What to notice
Outside the circle, rays near the direction of the centre meet it twice and the others miss it. For the conic with , Hardy’s convention () only gives points when , and they form one branch of a hyperbola. Allowing negative adds the other branch in the opposite colour: the same equation, a different curve.
Worked examples and exercises
Find the polar coordinates of , and , and the cartesian coordinates of the point with , .
Hint
Solution
: , and the point is in the third quadrant with , so . : , . : , . Finally and , so the point is . Note that alone would give the same answer for and ; the quadrant must be read off separately.
Find the cartesian equation of and the polar equation of the line .
Hint
Solution
Multiplying by : , that is , or : the circle with centre and radius . (Multiplying by can only add the origin, which is on the circle anyway.)
For the line: , which is . So , , and the polar equation is .
Let and . Show that the points satisfying with form one branch of the hyperbola , and find the polar equation (with ) of the other branch.
Hint
Hint
Solution
If with , then , so , and squaring gives , which is . Since this is a hyperbola, and it consists of the points with and those with . Our points satisfy , that is . Conversely, a point of the hyperbola with has and satisfies the polar equation. So the polar equation gives exactly the branch on which .
The points with satisfy , that is : the other branch. (Hardy’s footnote prints here, a sign slip: that equation gives , a branch of the mirror-image hyperbola . Check with , : the vertex of the other branch has , , and .) In the convention that allows negative , replacing by and by describes the same point, and the single equation then gives both branches.
For the circle with (origin outside), find for which directions there are two points, one point, or none.
Hint
Solution
. Real values need . The product of the two roots is and their sum is , so both are positive exactly when and both negative when . Hence: two points when and ; one point (a tangent) when and ; none otherwise. The rays that meet the circle are those within the angle of the direction of the centre.
Check yourself
What is for the point ?
. The sign of affects (here in the second quadrant), never .
In Hardy’s convention, the equation (with ) gives a point of the curve for every direction .
False. For the formula gives , which is not allowed; those rays miss the circle, which lies in the half-plane . At and the formula gives , the origin.
With required, the equation represents
Here and , so the curve is a hyperbola, but forces and gives one branch only (Hardy’s footnote). With negative allowed the same equation gives both branches. An ellipse needs and a parabola .
Which value of (with ) belongs to the point ?
The point is straight down from , which is three right angles anticlockwise from the positive axis of . The direction is the same direction, but Hardy takes between and ; and would point up.
Hardy’s language
Summary
- Polar coordinates: , , ; for Hardy and .
- A polar equation says, for each direction , how far out the points of the locus are; there may be none, one or several.
- A line at distance in direction : . A circle: ; through the origin, .
- Whether negative is allowed changes what an equation represents: with is one branch of a hyperbola for Hardy, both branches otherwise.