Some additional info about US subbands-by-mode, in reply to Klystron's
mention of exclusive Morse-code-only band segments.

In the following discussion, "modes" means "modes authorized for use
by amateurs on the specific amateur bands in question".

The current US regulation of the HF amateur bands permits Morse Code
everywhere, voice and image modes on specific subbands, and data modes
wherever voice is not permitted. Morse Code has no exclusive subbands
at all,
and is rarely used in the 'phone subbands. (I've been an active ham
40+ years and never used Morse Code in an HF voice subband). These
regulations are descendants of those going back many decades, to times
when amateur operation on HF consisted of Morse Code, voice and
nothing else. (For example, HF RTTY operation by US hams was first
authorized in the late 1940s, but only 45.45 baud 5 level Baudot code
was allowed.)

A few years ago, ARRL proposed "Regulation By Bandwidth", which would
have separated the various modes by
the bandwidth of the signal rather than whether it was voice, data,
image, etc. For example, under the proposal,
any mode less than 500 Hz wide would be allowed in the 500 Hz and
wider subbands, regardless of whether it carried
voice, data, image, Morse Code or other information. There were also
proposed changes to where automatic and semi-automatic data-mode
stations could operate.

The proposal got an RM number and a comment period. The comments from
those interested were overwhelmingly against the proposal. It was
revised but to no avail; ARRL finally withdrew the proposal.

IMHO, the most common reasons for opposition that I saw reading the
comments were these (in no particular order):

1) 'Phone operators did not want any data modes in the 'phone
subbands.
2) "Robot" (unattended) digital stations should be confined to small
subbands.
3) Concern that amateurs would have to be able to measure the actual
occupied bandwidth of their transmitted signals or be subject to
violation notices and complaints. Older equipment and hams who could
not afford spectrum analyzers would be forced off the air seemed to be
a common fear.
4) AM voice would be limited to 9 kHz bandwidth and was essentially
"grandfathered", but other modes could not
exceed 3.5 kHz on most bands
5) The existing rules did not need changing.

The FCC did act on an earlier "refarming" proposal by ARRL, and
widened the 'phone/image subbands on some of the HF bands at the end
of 2006. However, FCC went far beyond the ARRL recommendations in the
amount of change. This effectively reduced the spectrum space
available for data modes on those bands, since they could not be used
where 'phone is allowed. The most radical change was on the 80/75
meter bands.

About the same time as the "Regulation by Bandwidth" proposal, a group
of less than a dozen amateurs
calling itself the "Communications Think Tank" (CTT) proposed the even
more radical change of eliminating subbands-by-mode completely, and
simply specifying a maximum signal bandwidth for each band.

This proposal also got an RM number and a comment period, but the
comments were even more solidly against it than against "Regulation by
Bandwidth". The opposition was so overwhelming that CTT also withdrew
its proposal.

The point of all this is that ARRL and others have made proposals to
fundamentally change Part 97 in ways that would
favor the use of data modes, and the US amateur community has
repeatedly and strongly opposed those proposals.

73 de Jim, N2EY