Wavenumber range reversed in IR spectroscopy

I recently stumbled upon a result table from an IR spectroscopy, that is an image full of peaks. The goal was to decide which peak belongs to which group by looking up the intervals in a related table.

While the task itself is pretty straightforward, I can't explain to myself why the axis, as well as the intervals in the table, are "reversed" - meaning the higher values are on the left and the lower are on the right. In my opinion, this makes it much harder to quickly estimate if the given peak is in an interval or not:

e.g. for a number 2053, compare those two sets of intervals:

2100 - 2000
2070 - 2010
2050 - 2030


vs

2000 - 2100
2010 - 2070
2030 - 2050


The estimation seems to be easier with the second, at least for me.

Can someone provide a background on why the ordering is high-to-low and not low-to-high?

NOTE: I don't have any real background in the topic and so my view is rather naive.

• Spectroscopy is unfortunately littered with lots of different conventions. It's very common for the horizontal axis to go from high wavenumber to low wavenumber. If you do a Google image search for "IR spectrum of <your favourite chemical>" you'll see what I mean. So, "2100-2000" actually goes from left to right on the actual spectrum. – orthocresol Mar 6 '17 at 15:25
• It seems to come from the time when chart recorders were used, but so many spectra were recorded this way that when everything became computerised the convention was maintained. Its the same for nmr. – porphyrin Mar 6 '17 at 15:40
• @orthocresol I see, thanks. However, the question still holds - if the spectrum would go from low wavenumber to high, it would be easier to read the intervals, and I believe that's stronger than any other reason behind this. – Gyfis Mar 6 '17 at 15:52
• The convention of peaks pointing down in IR (when plotted as %T, which is customary for structure elucidation) also traces to the physical design of early instruments. I'm surprised that didn't catch your attention more than the reversed x-axis. – Bryan Hanson May 5 '17 at 22:44
• The peak convention does not make the task of reading the intervals harder, so that's why I didn't concern myself with it. – Gyfis May 25 '17 at 16:08