They are A BIT blurry
The one on the left, looks to be clone, the REAL package design was on the right...NOS
Both got wrapped into TO-92 style from the ISO conversion that makes China the big player it is now...no one else took the time to re-tool - they left their packages in old bins and warehouses to be gather up by collectors and repair personnel - the ones that tried to re-tool - never went back to make the OEM parts they used to - look at Motorola and Fairchild...
IT's (Right) "flared" for a reason these devices were originally designed to be recognized as small signal unique parts - so they did up various package designs before ISO standards arrived in the late 90's.
The "Flare" makes the part unique as in not to put parts in close proximity to package - hence the "lip" acts like a guard to keep components that could have their leads slide and touch the part - this was put on by various makers making FET or J-FET's - wanted to let others know of the output space on the board needed room to insert and not to have parts carrying signals close to the body, so it would not receive stray signals.
They also used these package designs to identify the kinds of substrate the part was, Germanium versus Silicon and even Selenium
But in your case...
View attachment 43260
Look at the "Letter" (A-B-C) designations.
These are showing High-gain "lot lettering" to show the level of Gain
Look on the bottom though.
The power dissipation (Idss) levels are what you need to be sure you select the part that operates
-
within if slightly above - your lowest expected number.
It's the Gain part, that may be your whole problem.
IF you're sure of your wiring, and can verify the tuning of L14 is correct - for L14's output not only captures the SSB signal, but offsets the Bridge to show RF - AND the IF is aligned correctly - the AM side of the Bridge and the "detuned" effect of L14 for AM (7.8MHz IF to 455kHz)) and SSB IF 10.7MHz - are NOT mixing together - throwing the leakage into the S-meter - you're issue may be gain - but when you work on a radio ,then the S- meter goes pegged, can't help but wonder; with the soldering iron hot, it touched something it shouldn't have.
Package:
This mess was during a transition between metal cased to Plastic epoxy - started in the 70's and as parts got obsoleted and NOS - they switched more and more to the epoxy case style but still had to keep the designs unique for identification of maker as well as proximity.
So the LEFT would be typical of the era, but the OEM part you're looking to use, is on the Right. Since they clone these - the Gain gives' away the effects of this cloning - you may have to reconsider finding a lower gain part or keeping the original in place until you're absolutely sure the FET was and is - truly damaged. As
@nomadradio said, they don't "fail" - it's usually something upstream that shows up further down the strip it goes...(a good tune will help go a long way in proving this theory you have)
They still make metal case, but for military spec and EHF designs requiring a shielded, not just isolated - design.