Coil Q, with form/without

Here's some results reguarding Q factors of the formless coils I've made 

Coil specs:
4" dia 7" long tight wound with 25 gauge enamel magent wire
2 coats of celluose nitrate based laquer

Form: 4" PVC (not baked) approx 1/16" wall thickness

The Q was estimated by finding the center frequency fo, and the 2 half 
power ( .707 amplitude) points, fo+df and fo-df. Then one gets Q=fo/(2df).

The coils were "base" driven with a signal generator and the loading of the 
50ohm generator impedance was adjusted for by keeping its output voltage 
fixed a some small value (1 volt). The ouput of the coil was monitored with 
an o'scope probe without actually attaching the probe to the top of the 
coil. The probe was mounted a few inches above the coil for very loose 
capacitive coupling through the air. (I didn't want to load the coil with 
the scope probe).

With PVC form:
fo=781.7 KHz
2df = 3.6 kHz
				Q= 220

Without PVC form:
fo=787.6 kHz
2df=3.3 kHz			Q= 240 

	%Difference = 9%

**********Suprisingly small difference!

Ferrite Chunk stuck in the middle of the coil (no PVC)
fo=758.8 kHz
2df= 3.2 kHz			Q= 237  lowered the frequency without
					changing Q!

Small cup (styrofoam) of water placed in the center of the coil (no PVC):
fo=772.5 kHz
2df= 3.5 kHz			Q= 220 

The next thing I'd like to do is get some different wire insulations to see 
how big an effect they have on Q.