To avoid confusion all species names should be written in full at first, and can then be abbreviated. In this particular example it would probably not be a problem to use the second, shorter version. But what if you had four species divided into three genera with all names starting with 'A'? As for question 2, the full name usually only needs to be used in the title (if relevant) and once in the body text (which includes the abstract), so you can abbreviate after the abstract. However, this is really only a space issue, and you can use the full name at all instances (you may abbreviate, but don't have to), unless of course journal style prohibits this. I find it useful to write the name in full both in the abstract and once in the main text.
See The International Code of Zoological Nomenclature (Appendix B: General Recommendations) for a reference. To be picky, technically, authorship and year of species names should also be included at least once in each paper (e.g. Amphiprion clarkii Bennett, 1830), to clarify exactly what definition of a name you are using. However, this is often not done in practice (in fields outside taxonomy/systematics). What you often see is that people refer to a authoritative source, e.g. "... all species names used in the study follow Doe et al. (2008)."