One example of such contamination is the pallavi of karunAsamudra where many sing Mmgrr,, in a style that almost matches the pallavi of O rAjIvAkSa. The phrase should rather be MG,,rr,,. However there are AlApanas by several stalwarts that are equally Arabhi and dEvagAndhAri!
I feel that the main culprit in this, however, is a fondness to oscillate the ris and dhas in Arabhi too.
I feel that Arabhi is neatly-playable on flute (maybe nAgaswaram) and on split-fingering-rich veena styles. Violin, I think, not that suited for it.
sAma can be instantly told apart from Arabhi and dEvagAndhAri by the fact that it doesn't use the niSAda.
I like avoiding ni in Arabhi too. Some versions of sAdhincene don't use it. Well, basically that becomes a ssd much like mgr is like mmr.
Dr PinAkapANi had mentioned to me that the first phrase in that set, srgsr,, is a sAma phrase, and has no business being in dEvagAndhAri. I feel that singing the first r flat or s dependent can tell apart the sAma and DG versions respectively.
srgsr is present in kShIrasAgarashayana too -- "tArakanAma tyAgarA
januta". I know many of the supposed "signature-of-the-rAgam krtis" face resistance, but I think kShIrasAgarashayana is a good one.
But if that, and the mgr I mentioned are all disputed, I'm beginning to see why dEvagAndhAri is a dEshya rAga.
And Keerthi, you forgot to criticize modern version of MD's annapUrNE. Which is the Swati Tirunal krti you're talking about btw?
Good that you mentioned the lingering ma of sAma. Probably one lingering feature in the modern sAma?