I’m stuck on what feels like it should be a classic locus/construction problem. Given a fixed segment AB and a fixed angle θ, I want the locus of all points P such that ∠APB = θ. I’m trying to do it with straightedge and compass only (no protractors, no measuring lengths numerically), and I keep going in circles.
Here’s what I think I understand: the locus should be two arcs through A and B, symmetric with respect to the perpendicular bisector of AB. For the special case θ = 90°, I’m fairly sure Thales’ theorem says the locus is the circle with diameter AB (and I guess that means A and B themselves are excluded because the angle would be 0° there?). But when θ is something like 45° or 110°, I don’t know how to pin down the exact arcs purely by construction.
My first attempt (probably wrong): I tried to force a triangle by setting the base angles. Since ∠APB = θ, the base angles at A and B would sum to 180° − θ. So I constructed at A a ray making angle (180° − θ)/2 with AB, and did the symmetric thing at B. Those two rays intersect at a single point P, which does give ∠APB = θ, but obviously that’s just one point, not the whole locus. So I realize I’m mixing up “find one triangle APB with apex angle θ” versus “find all such P.” This gives me only a single solution point per chosen side, not the arc.
Second attempt (feels closer but I don’t know how to finish it): I know that if P runs along the arc, then AB is a chord of some circle and the inscribed angle over that chord is θ. That means the central angle ∠AOB (where O is the center) should be 2θ. Also, O must lie on the perpendicular bisector of AB. So in principle, if I could place O on that bisector so that ∠AOB = 2θ, I’d have the right circle(s). But I’m not sure how to locate O without sneaking in trigonometry like R = |AB| / (2 sin θ). Is there a clean geometric construction that plants O at the correct distance along the bisector using only copying/combining angles and basic intersections? I tried to build an isosceles triangle AOB with vertex angle 2θ and base AB, but every version I tried ended up secretly using measurements or assuming lengths I can’t justify.
Concrete example: take AB as a 6 cm segment, and let θ = 45°. I can draw the perpendicular bisector easily, and I can draw a 90° central angle somewhere, but I don’t see how to “slide” the center up or down the bisector to make sure ∠AOB is exactly 90° while still keeping OA = OB = radius and AB as the chord. For θ = 110°, I get even more confused about which side(s) of AB the arcs live on and how to ensure I’m not accidentally drawing the supplementary arc.
Could someone explain a straightedge-and-compass method that actually constructs the two correct arcs through A and B for a given θ, and also why it works? If there’s a standard trick like reflecting something or using an auxiliary angle or a homothety, I’d love to see it step by step.
Follow-up questions:
– Do A and B ever belong to the locus, or are they always excluded because the angle at P would be 0°? What about the limiting cases θ → 0° or θ → 180° – does the locus collapse to a line or something degenerate?
– If I’m told “choose the arc on the same side of AB as some point X,” what’s the safest construction step to pick the correct side without relying on eyeballing?
– And bonus: if the condition changes to “∠APB ≥ θ,” is the locus then a closed region bounded by those arcs? If so, how would I construct just the boundary in that situation?
Sorry for the long-winded description – I’m mainly confused about how to locate the center(s) with only pure constructions, and how to be precise about which arcs I’m supposed to get.















