macro question

[sneakywumpus]

Can anyone explain to me this example from the DRI MAC manual?
This listing is generated by RMAC. MAC behaves the same.

                ;       MACRO PARAMETER EVALUATION
                ;
                MAC1    MACRO   A,B,C,D,S
                ;
                ;       ENTERING MACRO 1:
                        DB      '&A &B &C &D'
                        DB      S
                A:      NOP
                        MVI     B,1
                C&1:    NOP
                L&A&D:  NOP
                ;       LEAVING MACRO 1
                ;
                        ENDM
                ;
                MAC2    MACRO   E,F,G,H,S
                ;
                ;       ENTERING MACRO 2:
                        DB      '&E &F &G &H'
                        DB      S
                        MVI     M,H
                        MAC1    E,F&M,A,H,S
                ;       LEAVING MACRO 2
                        ENDM
                ;
 000F =         X       EQU     15
                        MAC2    I  ,,  X+1,  % X + 1, 'kwote'
     +          ;
     +          ;       ENTERING MACRO 2:
 0000+492020582B        DB      'I  X+1 16'
 0009+6B776F7465        DB      'kwote'
 000E+3610              MVI     M,16
     +                  MAC1    I,M,I,16,'kwote'

Why is the third parameter I? It should be A IMHO. To me it looks like the assembler is putting macro dummies in the global symbol table, and this references the A assignment to I in MAC1 from the MAC2 call from the first pass. Aren't macro dummies local to the macro? Is this a implementation feature/bug in MAC & RMAC?

     +          ;
     +          ;       ENTERING MACRO 1:
 0010+49204D2049        DB      'I M I 16'
 0018+6B776F7465        DB      'kwote'
 001D+00        I:      NOP
 001E+3601              MVI     M,1
 0020+00        I1:     NOP
 0021+00        LI16:   NOP
     +          ;       LEAVING MACRO 1
     +          ;
     +                  ENDM
     +          ;       LEAVING MACRO 2
     +                  ENDM
 0022                   END

[sneakywumpus]

The funny thing is the text that accompanies this example:

"Note that the concatenation of C with 1 reduces to a concatenation of A with 1 since C's value is A. The replacement of C by A constitutes
a substitution of a single occurrence of a dummy parameter, and thus the A which is produced is not itself replaced at this point."

So they say that C's value is A and not I. I think, I answered my own question :-), this is not the intended behavior.