Html conversion using Lynx has missing chars

My vm-8.2.0b setup uses "lynx" to read and convert MIME type text/html to display html messages inline.

This works fine when Emacs is run in a graphical frame. However, when I run Emacs in a terminal (e.g., under PuTTY), I get a lot of A-umlaut characters wherever there is a non-standard character (e.g., quotations, double space, etc.).

I was able to replicate this by passing the html part manually in a (non-Emacs) shell by piping the html onto lynx using:
   lynx -force_html -dump -stdin

I found that I could make the A-umlauts go away by adding the following switch to lynx:
    -assume_charset=utf-8

Now the interesting thing is that the MIME part does have:
   Content-type: text/plain; charset=utf-8
   Content-transfer-encoding: quoted-printable

But clearly the charset is not detected and/or passed on to lynx by vm-mime.el

I imagine that the reason that the display works fine in a graphical frame but not in a terminal is that graphical Emacs has stronger character fu logic to decode non-standard ASCII characters.

So, my questions are:
1. Is this a bug?
2. If not, am I doing something wrong? And what should I be doing to make the display work right in a terminal
3. More generally, should the lynx command line referenced in vm-mime.el be set to a variable (or alternatively should the switches be defined in a variable) so that one can twiddle with the exact switches sent to lynx.