The Library of Congress no longer considers Serbo-Croatian as a single language with 2 alphabets. Instead, Bosnian, Croatian, and Serbian are now treated as separate and distinct languages.
For Serbian, use the language code srp, regardless of the alphabet.
For Croatian, use the language code hrv, regardless of the alphabet.
For Bosnian, use the language code bos, regardless of alphabet.
For LC guidelines on how to distinguish these 3 languages, see "Guidelines for the
Identification of Serbian, Croatian, and Bosnian".
For Serbian language works, a 546 language note is routinely supplied because
it is necessary to designate the script of the work, since Serbian is commonly written in both the Cyrillic and roman alphabets. However, these notes are not routinely supplied for works in Croatian or Bosnian. Such a note for Serbian is normally formatted as below, if the note is only identifying Serbian as the language of an item:
546 In Serbian$b(Cyrillic).
546 In Serbian$b(roman).
However, if the identification of the alphabet does not end the field, then the tagging for the subfield is left off, e.g.
546 In Serbian (Cyrillic) with a summary in English.
For information on changes to the language codes for Croatian and Serbian, see: Language Codes for Croatian, Serbian, and Serbo-Croatian.