This dissertation presents a phonetic description of acoustic and perceptual properties of tones as well as acoustic properties of vowels in Meixian Hakka. The vowel and tone systems of Meixian Hakka have been described in the past studies. However, the basis of such descriptions is mainly individual impression and there is a considerable discrepancy in transcription among linguists. Due to the lack of experimental data presented in the literature, it is worthwhile to investigate the sound system of Meixian Hakka in acoustic and perceptual perspectives. The F0 patterns of the citation tones in Meixian Hakka were analyzed based on the raw and normalized data. It shows that there are acoustically six citation tones in Meixian Hakka. Yin Ping is a mid level tone and is transcribed as [33]; Yang Ping is a slight falling tone at the low frequency range and is transcribed as [11]; Shang Sheng is a mid-high to low falling tone and is transcribed as [41]; Qu Sheng is a high to low falling tone with a delayed falling contour and is transcribed as [51]; Yin Ru is a midhigh to low falling short tone and is transcribed as [41]; Yang Ru is a high level short tone and is transcribed as [55 or 5]. In spite of duration, the overall contour shape of Shang Sheng resembles that of Yin Ru. The two tones have a simple falling tone with an onset at the middle frequency range and an offset at the low frequency range. The comparison of F0 patterns of the citation tones on vowels [i, a, u] indicates that high vowels such as [i] and [u] tend to have a higher intrinsic fundamental frequency (iF0) than low vowels such as [a]. The correlation between F0 and vowel height is observed, not only in Meixian Hakka, but also in languages from different language families as described in the past studies and such correlation is held to be linguistic universal. There is a tendency for tones in CV(C) syllables to begin at a higher frequency than tones in V(C) syllables. Such difference in F0 pattern between CV(C) syllables and V(C) syllables may be associated with the effect of neighboring consonants on the F0 onsets of the following tones. Tone sandhi rules in Meixian Hakka are formulated in this research. The rules are to be compared with the tone sandhi rules in the past studies. The demonstration of F0 patterns of tones in two-tone combinations has shown that there are three sandhi rules in Meixian Hakka and they are as follows: (1) Yin Ping [33] is realized as a mid to high rising tone [35] when it is followed by [11, 41, 41]; (2) Shang Sheng [41] is realized as a mid level tone [33] when it is followed by [11, 41, 41]; and (3) Qu Sheng [51] is realized as a high level tone [55] when the second syllable is superimposed by [11, 41,51, 41]. Two sandhi forms from Yin Ping [33] and Qu Sheng [51] that are different from the six citation tones in Meixian Hakka are found. They are correspondingly the high rising tone [35] and the high level tone [55]. The tone sandhi patterns observed in this study are consistent with the patterns of Lü (2006) which were based on the acoustic data, but not with other past studies which were mainly based on individual impression. In addition to the acoustic characteristics of tones, the perceptual properties of tones and the correlation between perceptual and acoustic characteristics of tones in Meixian Hakka are evaluated. The potency of tonal height, slope, F0 onset and F0 offset in the identification of the four long citation tones has been examined. It shows that there is a strong correlation between tone perception and the acoustic characteristics of tones in Meixian Hakka. Tonal height plays an important role in the identification of the two level tones [33, 11]. A high F0 onset is an important cue for the identification of [51]. Slope is potent for the distinction between the three tones [51, 41, 11]. Stimuli with a steep falling contour are highly identified as [51], whereas stimuli with a less steep falling contour are identified as [41]. Stimuli are identified as [11] when the slope of falling is small. The perceptual results of this study show that not only slope but also tonal height are important for the distinction between [51] and [41]. The acoustic characteristics and temporal organization of monophthongs, diphthongs and triphthongs have been examined in Meixian Hakka. There are five Meixian Hakka vowels in the open syllable and they are transcribed as [i, e, a, o, u]. Results indicate that the relative distance between mid and high vowels differs with regard to vowel backness and speaker gender. The diphthongs in Meixian Hakka are suggested to be divided into two categories according to the degree of resemblance between formant patterns of the diphthong elements and formant patterns of their corresponding target vowels in the acoustic space. Such claim is supported by the temporal organization of diphthong. The results indicate that the dynamic nature of diphthong in terms of F2 range of change is sufficient for the distinction among diphthongs in a language such that the dynamic properties might be language-universal. In consideration of the phonetics and phonology of Meixian Hakka, it is proposed that there are 9 Meixian Hakka diphthongs and their narrow transcriptions are [ie, ia, io, iu, ai, Ɔi, Ʊi, ao, ƐƱ]. Besides, it is suggested that there are two triphthongs in Meixian Hakka and they are [iüi] and [iao].