If you are seeing two dips across the same band,
there are two likely causes: either it is due to
other resonances within the antenna, or due
to a resonance in the coax cable (common
mode current).
Some wideband antennas are designed to have
such a curve, particularly for 80/75m (3.5 to 4
MHz in IARU Region II). But there should be
no need for such behavior across the narrow
WARC bands.
The manual doesn’t give sample curves, so
we don’t know if this is considered normal.
In this antenna, the 17m and 12m bands
have adjacent elements that are close to
resonance: the 15m reflector will be tuned
close to 17m, and the 10m reflector close
to 12m. That can cause the response you
are seeing, depending on the exact tuning
of the parasitic element.
If the antenna doesn’t have an effective
balun, then the length of the coax shield
(and what is connected to it) can add
other resonances to the antenna SWR
curve. A quick check would be to add
3-4m of coax and see if one of the
SWR dips moves in frequency. I’ve
seen this a couple times when tuning
an 80m dipole: dips at 3.55 and 3.95
MHz, with one of them affected by
tuning the antenna and the other by
the length of the feedline. The solution
in that case is to add an effective balun
at the antenna feedpoint.