We show that it is possible to image the structure of a distant quasar
on scales of less than or similar to 1 AU by constructing a telescope
that uses a nearby dwarf star as its ''primary lens'' together with a
satellite-borne ''secondary.'' The image produced by the primary is m
agnified by similar to 10(6) in one direction but is contracted by 0.5
in the other and therefore contains highly degenerate one-dimensional
information about the two-dimensional source. We discuss various meth
ods for extracting information about the second dimension including ''
femtolens interferometry'' where one measures the interference between
different parts of the one-dimensional image with each other. Assumin
g that the satellite could be dispatched to a position along a star-qu
asar line of sight at a distance r from the Sun, the nearest available
dwarf-star primary is likely to be at similar to 30 pc (r/45 AU)(-2).
The secondary should consist of a one-dimensional array of mirrors ex
tending similar to 170 m to achieve 1 AU resolution or similar to 350
m to achieve much less than 1 AU resolution.