In the hyperbolic plane, suppose $L$ is a line with points $A, B, C$ on the same side of line $L$. Assume further points $P, Q, R$ are on line $L$ such that segments $AP$, $BQ$ and $CR$ are perpendicular to $L$ and the three lengths $AP$, $BQ$ and $CR$ are all equal. Prove $A, B, C$ are not collinear.

Suppose they were collinear. Then $ACRQ$ is a rectangle since all the angles are $90$ degrees due to being perpendicular to the line Rectangles do not exist outside of Euclidean geometry since the interior angle sum is less then $360$ degrees.

Would that be sufficient to show its a rectangle? Or is there a better approach to take. Or is there any direct proof ideas that could work

  • $\begingroup$ $\square ACRQ$ has right angles along $L$, but it's not obvious that collinearity of $A$, $B$, $C$ forces it to have more right angles. $\endgroup$ – Blue Apr 18 '18 at 20:56
  • $\begingroup$ How else could I show its a rectangle then? $\endgroup$ – HighSchool15 Apr 18 '18 at 21:42
  • $\begingroup$ Don't you want ACRP instead? ABQP and BCRQ are rectangles too. $\endgroup$ – Somos Apr 18 '18 at 22:22

We may assume $Q$ is between $P$ and $R$.


Claim. In $\square ABQP$, $\angle PAB \cong \angle QBA$.

As noted in the question, the angle-sum of a hyperbolic quadrilateral is less than four right angles, so the Claim's angles must each be less than one right angle.

Likewise, in $\square CRQB$, $\angle RCB \cong \angle QBC$, and these angles are each less than a right angle.

Then $\angle ABQ + \angle QBC$ is less than two right angles ---that is, the angles are not supplementary--- so that $\overline{AB}$ and $\overline{BC}$ cannot be collinear. $\square$

Proof of Claim. $$\triangle APQ \stackrel{\text{SAS}}{\cong} \triangle BQP \;\implies\; \overline{AQ}\cong\overline{BP} \;\implies\;\triangle PAB \stackrel{\text{SSS}}{\cong}\triangle QBA \;\implies\;\angle A\cong\angle B$$


I don’t see that the new angles you’ve created by drawing a line from $A$ to $C$ have to be right angles.

Here’s a rough argument that you may like, and may want to clean up:

Lemma. Let $\ell$ be a line in the hyperbolic plane, and $X,Y$ two points on the same side of $\ell$ and equidistant from $\ell$. Let $X',Y'$ be the points on $\ell$ that are gotten by dropping perpendiculars from $X$ and $Y$, respectively. Then in the quadrilateral $XYY'X'$, the angles at $X$ and $Y$ are equal (and less than $\pi/2$).

I think that to prove this, one may argue from symmetry somehow by exchanging $X$ and $Y$ as well as $X',Y'$.

Now let’s apply to your situation, assuming that your point $Q$ is between $P$ and $R$, and using your to-be-disproven hypothesis that $A,B,C$ are collinear.

In the quadrilateral $ABQP$, the angle $\angle ABQ$ is less than $\pi/2$ and in the quadrilateral $CBQR$, the angle $\angle CBQ$ is also less than $\pi/2$. But these angles have to be supplementary, since $B$ has been assumed to be on the segment $\overline{AC}$. No good


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.