Depending on your circuit, that 3.3M is as low as several hundred K instead. The pots in that "extra stage" are to mix the wet/dry signal. Properly done, it doesn't add much "gain" to the overall circuit.
There's different ways to add reverb and it depends on how much reverb you want/need too.
Most reverb (parallel side chains, verb is parallel with the main/dry signal) circuits have some signal attenuation for the main dry signal as John pointed to.
In this schemo link you posted there is no dry signal attenuation. I've never seen it done like that. The might have wanted the extra gain?
http://music-electronics-forum.com/attachments/2652d1214457798-5f2_reverb.gifThis link you posted is normal for adding verb with dry signal attenuation to balance the verb with the dry signal.
http://music-electronics-forum.com/attachments/6245d1252771722-5f2-reverb-6l6-el34.gifThe output signal from the verb tank is less than the output from a guitars PUP's. So 1 recovery gain stage after the verb tank will not give deep verb. The 2nd (verb) gain stage is called the 'reverb mixer' because it mixes the dry/wet signals together but it also boosts the wet signal back up to the dry signal.
In a Fender, -
The- bench mark for tube/spring tank reverb, the dry signal has 2 stages of gain from the guitar before the verb mix stage, the verb tank's output, less than a guitar PUP, has only 1 stage of gain before the verb mix stage.
They have to cut the dry signal so it doesn't swamp out the wet signal. But it's a balancing act of what amount of wet/dry signal you feed it.