I took the measurement reading you suggested and the reading was the same between all transistors with a variance of .1 volts.
The voltage in is +11.97vDC (+12vDC) @ Source and the voltage at the collector to ground is +11.70 give or take the .1 between the different types. That reading is with a 470 Ohm R on the 3 "odd balls" and a 1k R on all others.
I would still like an explanation to this problem even though I have solved it by switching to a ULN2803A. Knowing why the NPN setup was doing that will help in future projects I'm sure.
Sounds like you were not wiring up the transistors correctly? NPN transistor should be wired as a 'low-side' switch, with emitter lead grounded, base lead via resistor to arduino output pin, collector to load (LED negative) and LED positive to + external voltage source. External negative must also wire to arduino ground, as well as to emitter lead.
Is that how you wired it up?