Loanword adaptation provides important evidence for understanding how speakers modify foreign lexical forms to conform to the phonological structure of the recipient language. Differences between source and recipient languages may involve segmental inventories, syllable structures, phonotactic restrictions and permissible consonant sequences. When a borrowed form contains a structure that is not permitted in the recipient language, speakers may employ repair strategies such as vowel epenthesis, consonant deletion, substitution and assimilation. Among these strategies, vowel epenthesis is particularly significant because the quality of the inserted vowel is often systematic rather than arbitrary. The present paper reviews the literature on vowel epenthesis in loanword adaptation from the perspective of Feature Geometry. Particular attention is given to three mechanisms through which epenthetic-vowel quality may be determined: default vowel insertion, vowel-to-vowel feature spreading and consonant-to-vowel assimilation. Evidence from Shona, Sranan, Samoan, Kinyarwanda, Sesotho, Japanese and other languages demonstrates that these mechanisms operate differently across languages. The paper argues that Feature Geometry, particularly the model developed by Clements and Hume (1995), provides a useful representational framework for explaining these patterns because it treats phonological features as hierarchically organised and permits features or feature nodes to spread between segments. The review further considers directionality of feature spreading, contrastive feature specification, the relationship between phonological and phonetic explanations, and the interaction between default insertion and contextual assimilation. It concludes that vowel epenthesis should not be regarded simply as the mechanical insertion of an arbitrary vowel but as a structured phonological process involving syllable repair, feature representation, feature spreading and language-specific phonological requirements.