Both Sanskrit and Ancient Greek reflexes have a radical stress throughout the paradigm, which indicates that this accentuation was original, and that the PIE word for "brother" had an acrostatic inflection. Archaicity of the paradigm is furthermore indicated by the fact that the rest of the inherited kinship terms in Sanskrit have suffixal and desinential accent, so the pattern for the word for "brother" couldn't arise by analogy. The hysterokinetic inflection of other kinship terms (e.g. for *ph₂tḗr (“father”) and *dʰugh₂tḗr (“daughter”)) did exert influence in the two, which is evident by the absence of zero-grade in the suffix syllable (Sanskrit forms with -tā, -taram etc., Greek with -τηρ, -τερος etc.).
The root vowel is short in all of the reflexes, and no trace of lengthened grade can be found: in no language did Eichner's law operate (which predicts non-coloration of the *ē in **bʰrḗh₂-tr). Furthermore, in Balto-Slavic one can find an acute vowel (reflecting *bʰréh₂) instead of a circumflex vowel (which would reflect **bʰrḗh₂).
*bʰréh₂tēr m (oblique stem *bʰréh₂tr-)
- Lydian: 𐤡𐤭𐤠𐤱𐤭𐤳𐤦𐤳 (brafrsis) (brafrsis; stem brafr-)
- Balto-Slavic: *brāˀtē
- Celtic: *brātīr (see there for further descendants)
- Germanic: *brōþēr (see there for further descendants)
- Hellenic: *pʰrā́tēr
- Indo-Iranian: *bʰráHtā (see there for further descendants)
- Italic: *frātēr
- Mysian: βρατεραις (braterais, dat.pl.)
- Phrygian: βρατερε (bratere, dat.sg.)
- Tocharian: *procer < *bhrā́tēr