What sorcery is this?

So I have two images (which after import in unity) are of resolution as shown in the images. I’m not able to figure out why just one pixel makes such a difference in size of sprites after import.

**So just 1 pixel less in height makes such a huge difference. Is there anyone here who can throw some light on the import pipeline of sprites in Unity and how is this possible? Or what is causing this? What are the best practices? I have so many questions actually! **

P.S. Import settings are same for both sprites.

You’ve failed to notice the critical difference, which is that one image is ETC2 8bit compressed while the other is uncompressed RGBA.
I would guess that the extra pixel you’ve added contains alpha not supported by that compression format.

As a sidenote, you should always try to give your atlases POT dimensions to enable the most effective compression algorithms.