The topology of the observable Universe is not yet known .
The most significant observational sign of a non-trivial topology would be multiple images ( “ ghosts ” ) of a single object at ( in general ) different sky positions and redshifts .

It is pointed out that the previous search by [ Gott ] ( 1980 ) for ghost images of the Coma cluster can be extended by using highly X-ray luminous clusters of galaxies .
This is likely to be more efficient than with other astrophysical objects viewable on these scales since ( 1 ) X-ray clusters would be at least as easy to identify if viewed from other angles as any other objects and ( 2 ) the X-ray emitting thermally heated gas is likely to be simpler than for other objects .

Possibilities that the highly luminous cluster RXJ 1347.5-1145 ( z = 0 { \cdot } 45 ) has a “ ghost image ” at lower redshift are analysed .
It is noted that RXJ 1347.5-1145 , the Coma cluster and the cluster CL 09104+4109 form nearly a right angle ( \approx 88 ^ { \circ } ) with arms of nearly identical length ( 970 h ^ { -1 } and 960 h ^ { -1 } Mpc respectively ) for \Omega _ { 0 } = 1 , \lambda _ { 0 } = 0 curvature ( h \equiv H _ { 0 } / 100 km s ^ { -1 } Mpc ^ { -1 } ) .
This is a clue that the three clusters could be ghost images of one and the same cluster , for a hypertoroidal topology .
However , several arguments are presented that this relation is not physical .