I don't have my modelling program on this PC to look at such an antenna but it wouldn't be unreasonable to have only a useable 2 to 3 MHz bandwidth at most.
I have the spec sheet here for the Diamond X-50/200/300/500
antennas. The shortest X-50 has an SWR bandwidth at 1.5 : 1
of at least 6 MHz. As the gain increases with the longer versions,
the SWR curve gets sharper: for the longest X-500 model it
is about 2.5 MHz (but still less than 2 : 1 across the band).
That's as expected with the elements connected in series.
And it isn't uncommon to see some narrowing due to the
dual-band operation as well.
If you want a wide bandwidth, consider a stacked array of
folded dipole elements.
Otherwise, there are several options. One is to center it
in the part of the band where you do you do most of your
operating, or at least where you need the highest power.
You can see if the radio still shuts back as much when
you switch it to lower power - that may be perfectly
adequate for many local repeaters.
You can tune it for one end of the FM band and switch in
a matching circuit (likely just one coil or capacitor, or a
shorted piece of coax) to match it at the other end of
the band. Between the two you should be able to operate
the whole band.
It may be possible to design a matching circuit to go at
the antenna to increase the bandwidth without adding
significant losses, but that will depend on the actual
impedances measured at the antenna.