The curvature of spacetime produces non-euclidean behavior in light rays, and consequently multipathing and lensing effects. Einstein worked this out early - actually before he had the correct final form for the equations of General Relativity - but didn't publish it until 1936, and only then because he was pestered to by Rudi Mandl, a pesky Czech engineer. Einstein didn't consider the effect interesting or important.
Fritz Zwicky knew better and recognized the potential immediately (1937). Almost another half-century had to pass before experiment caught up with theory.
These are a few of the tidbits I have picked up in Evalyn Gates popular book Einstein's Telescope: The Hunt for Dark Matter and Dark Energy in the Universe.
Viewed through a normal telescope, a quasar looks like a point of light, much like a star. (Hence the name quasi-stellar object, which is abbreviated to quasar.) However, if a massive galaxy lies directly between the quasar and Earth, what we observe is stunningly different. The mass of the galaxy, which may be more than a trillion times the mass of the Sun, warps the spacetime around it and this warp acts as a lens that spreads the point of light from the quasar into a ring of light that appears to encircle the galaxy—the rare phenomenon known as an Einstein ring.
More often the quasar does not lie exactly behind the galaxy, but just slightly to one side of it. When this occurs, multiple images of the quasar can be produced—two or four identical copies of the same quasar are seen. This unmistakable hallmark of gravitational lensing was a known theoretical possibility in 1979, but the serendipitous discovery of a twinned quasar nevertheless took the observers by surprise.
Gates, Evalyn (2009-01-24). Einstein's Telescope: The Hunt for Dark Matter and Dark Energy in the Universe (p. 75). Norton. Kindle Edition.