Counter to intuition, the images of an extended galaxy lensed by a moving galaxy cluster should have slightly different spectra in any metric gravity theory. This is mainly for two reasons. One relies on the gravitational potential of a moving lens being time-dependent (the $\text{Moving}$ $\text{Cluster}$ $\text{Effect}$, $\text{MCE}$). The other is due to uneven magnification across the extended, rotating source (the $\text{Differential}$ $\text{Magnification}$ $\text{Effect}$, $\text{DME}$). The time delay between the images can also cause their redshifts to differ because of cosmological expansion. This Differential Expansion Effect is likely to be small. Using a simple model, we derive these effects from first principles. One application would be to the Bullet Cluster, whose large tangential velocity may be inconsistent with the $ΛCDM$ paradigm. This velocity can be estimated with complicated hydrodynamic models. Uncertainties with such models can be avoided using the MCE. We argue that the MCE should be observable with ALMA. However, such measurements can be corrupted by the DME if typical spiral galaxies are used as sources. Fortunately, we find that if detailed spectral line profiles were available, then the DME and MCE could be distinguished. It might also be feasible to calculate how much the DME should affect the mean redshift of each image. Resolved observations of the source would be required to do this accurately. The DME is of order the source angular size divided by the Einstein radius times the redshift variation across the source. Thus, it mostly affects nearly edge-on spiral galaxies in certain orientations. This suggests that observers should reduce the DME by careful choice of target, a possibility we discuss in some detail.

15 pages, 8 figures, 2 tables. This is the peer-reviewed version which has been accepted for publication in Monthly Notices of the Royal Astronomical Society