Thank you for your report. You are right there are some inconsistencies with this tag. Actually I don't know how to treat MS$FOCUSED_ION: ION_TYPE and I can not exactly find out from historical records. We have two very similar tags MS$FOCUSED_ION: ION_TYPE where I don't know what it should be used for and MS$FOCUSED_ION: PRECURSOR_TYPE which is clearly the type of the precursor ion in MSn spectra. Its exactly used for this purpose in our records. I guess its a good idea to start a discussion about the meaning and usage of MS$FOCUSED_ION: ION_TYPE here. Any comments would be appreciated.

And here are some numbers. We have:
37466 records with MS$FOCUSED_ION: PRECURSOR_TYPE

• 682 MS
• 35715 MS2
• 999 MS3
• 70 MS4

12240 records with MS$FOCUSED_ION: ION_TYPE

• 11972 MS
• 268 MS2

6 records with MS$FOCUSED_ION: ION_TYPE & MS$FOCUSED_ION: PRECURSOR_TYPE

A while ago e removed some MS$FOCUSED_ION: ION_TYPE from a number of record files if they had MS$FOCUSED_ION: ION_TYPE and MS$FOCUSED_ION: PRECURSOR_TYPE and both tags had the same ion and if they were MSn spectra.

The documentation gives examples which may help you see the difference - there are likely very few records that make use of this (backed-up by your numbers):
https://github.com/MassBank/MassBank-web/blob/master/Documentation/MassBankRecordFormat.md#251-subtag-ion_type

https://github.com/MassBank/MassBank-web/blob/master/Documentation/MassBankRecordFormat.md#251-subtag-precursor_type
and especially look at the MS2 vs MS3 example for that latter link, the MS3 one has more detail, this must be the field they defined to put in the ion relationships to rebuild the "tree", not just list the "immediate precursor".
It would be interesting to see what they did for both fields for some of the MS3 and MS4 (if they exist with these fields)?

Currently tag 2.5.1 structure is:
2.5.1 --- FOCUSED_ION (Information of Precursor or Molecular Ion.)
---------- subtag: PRECURSOR_TYPE (Type of Precursor Ion in MSn spectrum.)
---------- subtag: ION_TYPE (Type of Focused Ion.)

As my understanding, ION_TYPE is more appropriate for MS1 description. But in Tendem Mass Spectrometry, as show MS2 MS3, one ion type cannot describe multi precursor. Here, ION_TYPE is unsuitable.
In this case, is ION_TYPE more like "last precursor's type"?

Agree that PRECURSOR_TYPE seems the more appropriate and it's the one used in RMassBank. Should we update the record specifications to indicate that PRECURSOR_TYPE is preferred?