John and Mike,
I agree that heatsinks are prudent for the amp as it will contribute to the mosfets running in a more stable "on-resistance" area of the curve.
When I designed the PCB and tutorial, I took this into consideration, but decided to omit the heatsinks for three reasons.
* The amplifier is designed for brand new builders, some of whom have very limited electronics experience. As you know the tabs of the Mosfets are at B+ and the tabs of the regulators are at a +vdc. I wanted to avoid the chance of someone accidentally shorting the tabs out to a heatsink, or ground. Instant smoke on first power up is very discouraging.
* I ran some calculations and decided to run the class A current at .125 @ 12v. This will dissipate about 1.5w max, which is barely under the max dissipation to ambiant air (~1.6w +/- 20%). The mosfet itself drops about 6v @.125A so about 750mW
After a good month with the unit on I have not experienced any thermal runaway.
* The distance between the semiconductors did not leave enough space for off-the-shelf heatsinks.
Now for those who have successfully built a working unit, and desire to lengthen the life of their semiconductors, heatsinking will be a great upgrade to the amp. I recommend that each TO-22 package be heatsinked individually, as this will avoid the usage of insulators and nylon washers. I have seen some folks use coins, washers, and pretty much any metal as heatsinks. For this small amount any clip-on type or small piece of tin or aluminum would do the job. Just make sure that there are no shorts to ground or other components before power up. You can see that I fashioned some heatsinks on my internal amp because it runs B+ at 18V. The large heatsink is a 7812 for heaters (now that little guy can get hot, so I sinked to the chassis as well)

Thanks for the input
Rogers