All this assumes that the 25KHz receiver is precisely tuned to the exact frequency shown on the display. A bit of tolerance taking it closer to the adjacent transmission would make it much more likely to have break through.
In practical terms I agree, it’s a non-issue
Peter is right again… both receiver bandwith (receiver selectivity) is lower then channel spacing, and the occupied bandwith from the transmitted modulated signal is (far) less then channel spacing.
One must note that 60dB reduction in signal at the channel spacing edge is an awfull lot.
My brain was wired by communists 
My take on this is that we either have the “much debated” interference issue, or we don’t.
I don’t think we have unless you are talking about really old shagged radios, because the TX carrier deviation on AM is simply equal to the audio bandwidth, which is pretty well limited to about 3kHz.
So this reg allows us to keep the 2nd (25kHz) radio for 25kHz frequencies. These will always be the vast majority of GA-relevant frequencies, because we don’t have the stupid frequency tax (the UK abandoned that nutty-as-fruitcake proposal, last time I heard) so there is no incentive for stations to change their frequency.
8.33kHz merely introduces two more frequencies between each adjacent pair of 25kHz ones. This is obvious when you play with e.g. a KX165A radio.
Peter wrote:
I don’t think we have unless you are talking about really old shagged radios, because the TX carrier deviation on AM is simply equal to the audio bandwidth, which is pretty well limited to about 3kHz.
That much is clear, but what about the transmitter carrier frequency tolerance? According to ICAO Annex 10, the tolerance must be at most ±0.003 % for a 25 kHz channel but ±0.0005 % (six times better) for a 8.33 kHz channel. Can we rely on (reasonably) modern 25 kHz equipment to meet the 8.33 kHz requirement?
Peter wrote:
I don’t think we have unless you are talking about really old shagged radios, because the TX carrier deviation on AM is simply equal to the audio bandwidth, which is pretty well limited to about 3kHz.
Hmm… That’s right. Then I really wonder about the rationale behind making XXX.000 and XXX.005 distinct. Peter, do you happen to have a VHF spectrum analyser to look at the actual emission spectrum of your radios?
We must have wires crossed (not saying I am right) but AIUI the way 8.33 works is that say you have
130.000
130.025
currently.
With 8.33 you generate two more which are 8.33-spaced i.e you now have
130.000
130.0083333333333333333333333333333…
130.0166666666666666666666666666666…
130.025
but in an IMHO misguided attempt to keep pilots’ life simple, the above frequencies are not going to be displayed like that. Yhey will be displayed (and published) as
130.000
130.005
130.015
130.025
(from memory, playing with the KX165A)
They should display and publish them as 130.008 and 130.016…
There is no 130.005 in reality.
Therefore, there is no issue at all mixing 25 and 8.33 because 130.000 and 130.025 are still there as before. The only possible issue is with carrier frequency initial calibration tolerance and stability (which as I say should not be an issue in the latter 20th century, with 1ppm TXCOs costing peanuts), and possibly (this could be a bigger one) receiver bandwidth which if too wide could cause adjacent channel interference at the receiver. We did a thread on this here and here
I do have a spectrum analyser (10kHz to 2.4GHz) but it’s not a trivial job to bring it to the plane. However what would be the point? 130.005 cannot exist because it would be too close to 130.000 which will clearly remain.
Let’s stick with 130.000. If it’s a 25 kHz spacing assignment, the designator will be 130.000. That’s what you’ll see on the radio and on your charts. However, if it’s a 8.33 assignment, the same frequency would be designated as 130.005. So it’s:
130.000 – 130.000 MHz with 25 kHz spacing
130.005 – 130.000 MHz with 8.33 kHz spacing
130.010 – 130.008 MHz with 8.33 kHz spacing
130.015 – 130.017 MHz with 8.33 kHz spacing
130.025 – 130.025 MHz with 25 kHz spacing
You shouldn’t be experiencing interference because 130.010 shouldn’t be assigned as long as 130.000 is assigned. They create 8.33 kHz frequency assignments by decommissioning 25 kHz frequency assignments. That’s how I understand it.
130.000 is 130.0000 with a 25kHz channel spacing
130.005 is also 130.0000 but with an 8.33 kHz channel spacing
130.010 is 130.0083 with an 8.33 kHz channel spacing
130.015 is 130.0166 with an 8.33 kHz channel spacing
With 8.33 kHz, you don’t dial frequencies on the radio, you dial channels. As Martin and stevelup have written, 130.000 and 130.005 are different channels, but the same frequency. The difference between them is the frequency tolerance. If the radio is a 8.33 kHz set, then in practise the difference will be that with channel 130.000 the reciever will have a wider frequency tolerance.
