That new transistor is 2N5089, perfect for the project. No problem there.
The caps ..... their effects are still beyond my grasp. I hope someone else can chime in on that topic. But if you put two 22uF in parallel (side-by-side) it will equal 44uF which should be close enough to 47uF for the gain pot. If you take those same two 22uF caps and put them series (back-to-back) you would arrive at 11uF, which should suffice for the 10uF output cap.
As for the resistors, the 56K (due to 10% tolerance) could be anywhere from 50.4K - 61.6K
The 47K would have been anywhere from 43.3K - 51.7K
The two resistors have a small window of overlap at their extreme tolerances only. For a lot of instances this would be fine, but the 47K is part of the bias network and
might be the problem.
You could do some resistor math to get you close to 47K. If you have more 100K's hanging about, you could put two in parallel (side-by-side) to get you to 50K which is closer. Now if you check them with your multimeter and find two that are (due to tolerance) actually somewhat under 100K, then that would bring you even closer to 47K. Two @ 94K would be bang on 47K (in parallel).
Below is a link to a website with online calculators to figure out resistors/capacitors in series or parallel, as well as a program you can download to do it offline. You can even use it to identify resistor/capacitor values based on the markings on the component. Very handy!
http://www.electronics2000.co.uk/I hope that helps!