Power dividers.

My plan was to use 4 to 1 1/2 lambda power dividers that was 5 times 1/4 lambda on each side.
But for some reason this seems difficult.
Built my two dividers and they both behaves with less matching performance than on the 1/4 lambda resonance and 3 !/4 and even 7 1/4 lamba is good but not 5.
Measured other dividers of the same kind I built before and they more or less behaves in the same way.
So I decided to change to 3 1/4 lambda on each side instead and then it worked as expected.
Anyone seen this phenomena before?

Picture above shows measurements of the 3 1/4 lambda divider which is the same as a two meter divider.
The left resonance is for 432 MHz which here is the 2 nd resonance but the 144 MHz cant bee seen here as the mini VNA with extender cant work below 200 MHz.
So all is ok on 432 at the marker but look at the third resonace at approximately 700 MHz what a poor matching.
This is the 5 1/4 lambda frequency for that divider and close to 1GHz is the 7 1/4 lambda frequency which again is ok.
Its not a big concern but it will require some longer feedeing cables from the antennas to the divider but I am a bit puzzled and irritated about the thing.