Maybe after Dave's (G3WCB's) pertinent comment I should amplify my comment a
bit. I see no reason why we cannot have some agreements about useful
frequencies particularly whilst there are ways of making "stock" crystals
work in the band. This might be a good idea for stable data modes. What I
would like to avoid is some arbitrary arangement now, while the band is
still new, getting set in concrete and then vigoorously defended by
different factions so that it becomes impossible to work DX out of the
"approved" sector without being vilified for breaking "the bandplan". This
band will eventually exist across a number of countries who will have
different priorities to our own. At present we need to operate and
experiment and to use normal amateur good manners. As the band usage ( and
the operators ) mature, we will be able to translate our ad hoc arrangements
into someting more of "guide" to prefered procedure. There was never, and is
not now, any actual "bandplan" on 136 but you would never believe it ! An
example of this need for pragmatism is the use of 136.2 for east to west
Transatlantic QRSS....right in the middle of the "CW" section.....but this
was requested by the US stations as their best option to avoid CFH on 137
and the Dixon Navy station on 135.9.

So let's agree how best to use the allocation, but for heaven's sake DONT
call it a bandplan !!

On another front if anyone has worked out ways of using divisors of standard
cheapie crystals, please post them here. This a quick way to get on
initially. My  contribution is to note that the little block oscillators ( 8
and 14 pin DIL footprint sized ) can be shifted slightly in frequency
without affecting the stability too badly by varying the supply voltage.
Most of the nominal 5v version will start at about 3.5v and will run up to
about 6v, though they are starting to get hot there and thermal drift may be
more of a problem that at the lower voltage. I have used this technique for
VCXOing them. Some have an inhibit pin that could even be used for keying,
though with switching PAs that may lead to "unsociable" clicks. The shift
achieved could certainly be enough for DFCW, unless a very large divisor was
used. Also one of the standard easily available ceramic resonators is for
500kHz, others at 1000kHz, these will pull quite nicely, and a lot further
than quartz crystals will pull (see back issues of G3VA's TT in Radcom). The
1000kHz one would probably make a simple VCCO to cover the whole band, ready
for the squaring divide-by-2 stage.

Cheers de Alan G3NYK