Des échelles à sept intervalles équidistants n'étaient connues des ethnomusicologues, il y a encore peu de temps, que dans deux grandes régions: Asie du Sud-Est et Afrique noire (cf. Stumpf, 1901; Hornbostel, 1911; Tracey, 1948; Duriyanga, 1948; Jones, 1964; Morton, 1968; Rouget et Schwarz, 1969). Récemment, l'existence de cette échelle a été signalée chez les Indiens Tepehua du Mexique. Notre but est d'apporter du matériel nouveau pour une quatrième région du monde, la Mélanésie, dont on ne pensait pas jusqu'à présent qu'elle pouvait avoir des échelles de ce type.

Introduction. Until recently, equal interval, seven tone scales were known to ethnomusicologists only in two major areas: South-east Asia and Black Africa. However, not long ago this scale was reported in Mexico amongst the Tepehua Indians. Our aim here is to show its existence in yet a fourth area, Melanesia, which until now was not thought to have this scale. The article offers the initial findings from a continuing study into the scales of more than 300 instruments belonging to about 40 panpipe ensembles which we have recorded in different societies of Malaita, the most densely populated of the British Solomon Islands. In this initial report we have chosen only 28 instruments belonging to four panpipe ensembles of the ‘Are'are people: ‘au tahana, ‘au keto, ‘au taka iori and ‘au paina.1.

1. Recording and measurement procedure. The instruments used here for the scale analysis are the ones which appear in the recorded pieces from the two discs Melanesian Panpipes, ‘Are'are Vol. 1 and Vol. 2. The scales of each instrument was played in ascending direction by a musician, each note being blown six times, giving a total duration of about 5 seconds. A Stroboconn was used in the measurements. The degree of error in the readings of the pitches may be considered to be ± 3 Cents in the middle two octaves (C4- B5) and ± 5 Cents or the very low and very high registers. To ascertain whether there was a difference between the pitches when played in the scale order and when played in a performance, we measured the pitches of two pieces of music played with the three big instruments of the ‘au keto ensemble. These measurements were made from recordings of three part polyphony in which the three musicians, each in turn, played near the microphone. Each measurement was taken twice, and the comparisons showed that the difference between ensembles and the played scales was minimal, in the order of ± 5 Cents (cf. Table VIII).

2. ‘Au tahana ensemble. This consists of four instruments of two different sizes. The musicians play a two-part polyphony, each part doubled at the octave. Tabe I: Pitch measurement and interval sizes (each horizontal line represents one instrument under which are the Stroboconn pitch measurements; figures in italics above each line show the interval in Cents); Table II: Second interval discrepancies; Table III: Second interval discrepancies arranged in decreasing value; Table IV: Unison discrepancies; Table V: Octave discrepancies; Table VI: Pitch discrepancies between the four instruments.

3. ‘Au keto ensemble. This ensemble consists of six instruments of different sizes. The musicians play a three-part polyphony, each part doubled at the octave. Table VII: Pitch measurements and interval sizes; Table VIII: Comparative interval measurements; Table IX: Third interval discrepancies; Table X: Second interval measurements and discrepancies (in brackets); Table XI: Octave discrepancies.

4. ‘Au taka'iori ensemble. This ensemble consists of ten instruments of different sizes. The musicians play a four-part polyphony, two parts being tripled at the octave, and two being doubled at the octave; Table XII: Pitch measurements and interval sizes; Table XIII: Third interval discrepancies; Table XIV: Octave discrepancies.

5. ‘Au paina ensemble. This consists of eight instruments of four different sizes. The musicians play a two-part polyphony, both voices being quadrupled at the octave. Table XV: Pitch measurements and interval sizes; Table XVI: Unison discrepancies; Table XVII: Octave discrepancies.

6. Vernacular terminology. The ‘Are'are people have no names for the different tubes of the panpipes used for playing in concert, nor for the pitches obtained by blowing these tubes. However, they have names for the low (pau or kou) and high (kikiri) registers of an instrument. Two intervals have names: the second (rapi ‘au) and the octave (suri ‘au). Two other terms indicate the direction — ascending or descending — of an interval or melodic fragment; the term “to ascend” (hiuru) indicates to go towards a lower pitch, whilst “to descend” (hihu'a) indicates to go towards a higher pitch.

7. • The four types of panpipe ensembles can be distinquished by several features (cf. Table XVIII): (1) the number of instruments; (2) the number of tubes in each instrument; (3) the scales with conjunct degrees (seconds), with disjunct degrees (thirds), and with both; (4) the doubling, tripling, or quadrupling at the octave; (5) the range of each instrument and of the ensemble.

Three of the four ensembles described above show clearly the presence of an equiheptaphonic scale: the ‘au tahana (conjunct degrees), the ‘au keto and ‘au taka'iori (disjunct degrees). The instruments from the ‘au tahana ensemble studied here conform most closely to the theoretical scale (half the intervals are within 5 Cents, whilst the only intervals [6 of the 52] which deviate more than 20 Cents are found in the high register of the small instruments). The (au tahana ensemble is considered by the ‘Are'are as being the oldest and most important panpipe ensemble, whereas the others are considered to be derived. After an examination of both the terminology and the musical practice of the ‘ Are'are (there is no concept of modes; the same piece can start on different pitches), we feel that the differences in interval sizes shown in the analysis represent non-pertinent discrepancies in the equiheptatonic scale.

The ‘au paina is the only panpipe ensemble in which the ‘Are'are distinguish between different intervals: the second and the third. However, the scale analysis has shown that there is a difference in the sizes of the seconds; this is contradictory in view of both the oral tradition which claims that the ‘au paina was made from the ‘au tahana model, and also of information we received which suggests that it would be made today in the same way. During our next field trip we will endeavour to see how the instrument maker modifies the tuning for the ‘au paina from the ‘au tahana model.

The unison discrepancies between instruments of the same name and size (in the ‘au tahana and ‘au paina ensembles), unlike some Mandinka xylophones and the Balinese gamelan, do not show systematic characteristics. Contrary to the Balinese gamelan the ‘Are'are panpipe ensembles tend to have “stretched” octaves in the low registers and “compressed” octaves in the high registers.