"max N" clause on "pseudo-device pty" in config file

[Mark Rosenthal]

We're running Ultrix 3.0 on a VAX 8650.  Most people log in via ethernet.
We've recently added some X-servers, and now that users can open several
windows at once, the number of pseudo-ttys in use at any one time has
increased.  We recently ran out of pseudo-ttys, so we created more special
files in /dev, and added the appropriate lines to /etc/ttys.  I thought it
was probably necessary to allocate more kernel space as well, so I checked
in the Ultrix-32 documentation, System Management Volume 2.  The manual is
entitled Ultrix-32 Guide to System Configuration File Maintenance.  In section
"1.3.4 Pseudodevice Definitions" on page 1-16, it states:

    "Each pseudodevice definition line in the config file defines a driver
     for a particular pseudodevice.  Each pseudodevice definition line begins
     with the keyword pseudodevice, followed by the pseudodevice name.  The
     format is:
	 pseudo-device	name	[max n]
     The name is the name of the pseudodevice.  Configuration files can have
     the following pseudodevice names:

     pty	For pseudoterminal support (default = 16, specify max n for
		 more than 16)."

So, thinking no harm, I modified the config file to read:

    pseudo-device 	pty	max 64

and ran config, which promptly complained about the syntax of the line
immediately preceding the one I had modified.

It also occurs to me that we have been running fine with 32 pseudo-ttys
(tty[pq][0-f]), and if the documentation is to be believed, we should long
ago have run into a limit when we exceeded 16 pseudo-ttys.

So, what gives?  Does the kernel allocate them dynamically as needed?
Am I doing something wrong, or is the documentation just plain wrong?
-- 
	Mark of the Valley of Roses
	...!bbn.com!aoa!mbr