Has anyone ever seen/used germanium transistors in a TO-92 case?  TO-92 being the case that silicon transistors usually come in (black, one side round, the other flat.)

Nope. The germanium transistor died (to the industry that made them) before plastic cases became common. I don't believe any were ever made.

What about new production?  I got a bunch of transistors marked "B494" dirt cheap.  Salesperson insisted they were 2SB494s, which I've used with success before, only these are in TO-92 cases, PNP, ECB pinout.  Tested the gains with a cheapo meter in the store, 5 out of 5 were 130-140 Hfe.  Spent the dollar for a couple dozen of them, brought 'em home and tested them again - no leakage, no gain craziness when heated.  So my first thought is they're some other low gain Si PNPs which have been remarked - any idea what they could really be?

Interesting if they do turn out to be new production 2SB494's.

To the best of my current knowledge, the only new production germanium is a few companies that produce mainly for the military replacement market at quite high prices and also turn out a few special order devices like Dunlop's "NKT275" and the current production AC128's. Those are all in metal cans.

Here's what you want to test. Use a multimeter with a "diode" range on the ohmmeter scales. Measure a known-silicon diode like a 1N4148 or 1N4000 series, or a known-silicon transistor like a 2N3904. Then measure all the leads of the "B494" with respect to each other. The silicon junctions will read some value on the ohmmeter, often 6xx or 14xx. A real germanium junction will measure less than half that.

You can also use a 10K resistor and a 9V battery in series with the junctions and measure their voltage directly. Silicon will be 0.5V to 0.7V, germanium will measure 0.1 to 0.3V.

Good news for effects builders if it really pans out.

LONG AGO I BOUGHT A FIFTEEN pACK OF TRANSISTORS MARKED ON THE BOX AS GERMANIUM.
 THEY HAD LiGHT FINE MARKINGS< OF WHICH I DIDN'T EVER TAKE MUCH NOTICE EXCEpT THEY LOOKED THE SAME AS REGULAR SI< EXCEpT FOR THE LaCK OD OpAQUENESS OF THE MARKINGS< THEY ARE PNP>

hey petemoore? is you keyboard stuck? thats the first time I've seen peeps type like that....

Anyway, Look at the germanium transistors in the sale/trade(by me) those are what they can look like. All other germaniums are metal, except the very rare one thats in a plastic dome. The "triangle" case (TO-92) came into existence after germaniums were out, and even after many si trannies were made.

Rats.  Going by RG's helpful post, it looks like the "B494"s are NOT Ge.  Voltage drop between junctions is around .6V, which puts them into Si territory.  I wonder who the heck made these - if they were designed to be Si replacements for the real Ge 2SB494s, they're way off since the pinout is wrong (ECB instead of EBC).  If they're mislabelled, what are they really, since the only references to PNP "B494" I can find all point to Ge devices (I only did a *B494* cross on NTE's site).  

I suppose if you stretch out the middle C leg it could be made to fit into a TO-1 pinout without too much difficulty.  So my 2nd question is, are Si replacements for Ge devices common?  I guess this is more of a "back in time" type of question - what was the situation 30+ years ago when Si devices became popular, were they simply subbed into the same circuits with minor tweaks (like the FF?)

not ultra common, especially in audio, as you have to have in mind what kind of transistors your going to use when you design an amplifier. Also, Ge trannies have certain attributes to them that give the signal a "different" sound. To rough it up, for an audio circuit, a lot of work still is needed to be done to the si replacements at this point to make them sound like the ge ones.