Data Segment in Medium Model

[higgins at garnet.berkeley.edu]

   I am porting a Medium Model Microsoft 'C' Dos Application to
SCO XENIX 2.2.3 on an 80286. I am running out of heap space in the 
Data Segment under XENIX, while in DOS, with the same source code I have
plenty of room in the Data Segment for the heap. My Stack size is the
same in both environments and my compile and link switches are relatively
equal. I have the following questions regarding this problem.
   ( 1 ) Do I get the total 64k less the stack in the data segment for
         Xenix?
   ( 2 ) Are there any storage allocation differences between Microsoft 'C'
         on Dos and the 'C' compiler on Xenix?
   ( 3 ) What about the symbol table?

   I am sincerely grateful for any responses to this perplexing situation.
                                               John Tyler,
                                            The Higgins Group