Results of two listening tests aimed to find out how native listeners identify vowels from connected speech without the support of meaning.