The study of Portuguese fricatives is a complex problem, which has not been explored fully. As part of a larger study of the acoustic properties of Portuguese fricatives, a corpus based on 154 Portuguese words containing the fricatives /f, v, s, z, ʃ, ʒ/ in combination with the non-nasal vowels /i, ɨ, e, ɛ, ɐ, a, ɔ, o, u/ has been recorded. A nonsense word corpus was also devised consisting of /pVCV/ sequences. Sustained Portuguese fricative consonants were also recorded. Both the acoustic speech signal and the laryngograph signal were recorded in a sound treated booth. Observation of laryngograph and acoustic signals of Portuguese words has revealed that there is a large number of devoiced examples. Averaged power spectra were computed for all fricatives. For the nonsense word corpus, ensemble averaging based on one DFT computed at the same event in each of nine tokens was used. For all other corpora, time averaging of DFTs computed from nine 9 ms windows was used. Substantial differences are found between spectra of voiced and unvoiced, same-place fricatives. Other comparisons of effort level in sustained fricatives to position in the word add to a knowledge of the interaction of voice and frication source in Portuguese. The influence of particular vowel contexts in fricative devoicing is currently being investigated.