Hi A,
FvM is correct on both issues. Changing the load capacitor can be useful for moving the frequency a few hundred ppm. The direction in which it changes and amount in each direction will depend on the load capacitance in the circuit you have. Changing the load capacitor has the undesirable effect of changing the oscillator negative resistance - which can lead to no oscillation.
The only way I can think of to get 4.26 MHz or 4.38 MHz is to use two separate crystal units tuned to the appropriate frequency (allowing for the load capacitance of the oscillator). However, I noticed that the circuit note you provided details a mixer following the crystal oscillator in order to mix two tones and filter a harmonic. If that is the case, you might be able to mix the tones obtained using a 4.00,4.19 and 4.93MHz crystal unit with an appropriate second tone, filter out the undesired harmonics, and get 4.26 MHz or 4.38 MHz. For example, if you mix 4.00 MHz with 8.26 MHz, you will get tones at 12.26 MHz and 4.26 MHz...just a thought based on my reading