CL compiled, but I don't think it should have

Hi all,

I have a CL program that was causing a very strange error.  The CL has 
this at the beginning:
PGM        PARM(&VTYPE &PTERM &UTERM &DAYS1)

DCL        VAR(&VTYPE) TYPE(*CHAR) LEN(3)
DCL        VAR(&PTERM) TYPE(*DEC) LEN(3)
DCL        VAR(&DAYS1) TYPE(*DEC) LEN(3 0)
DCL        VAR(&UTERM) TYPE(*DEC) LEN(3)


Depending on which location the report was requested from, there was a 
either

CALL       PGM(pgmA) PARM(&VTYPE &UTERM &DAYS1),

which worked just fine, or

CALL       PGM(pgmA) PARM(&VTYPE &UTERM DAYS1)

which doesn't work.

My question is, how did the value DAYS1 pass through the compiler.  This 
is not a valid variable reference (no & prefix), nor is it a valid 
character literal (not enclosed in ' ').  Can anyone explain to me why the 
compiler didn't catch this?

TIA,
Adam

#####################################################################################
Attention:
The above message and/or attachment(s) is private and confidential and is 
intended 
only for the people for which it is addressed. If you are not named in the 
address 
fields, ignore the contents and delete all the material. Thank you. Have a nice 
day.

For more information on email virus scanning, security and content
management, please contact administrator@xxxxxxxxxxxx
#####################################################################################