388 Mr. A. J. Ellis .
TonometriealObservations on [ Nov. 20 II .
" Tonometrieal Observations on some existing Non-harmonic Musical Scales .
" By Alexander J. Ellis , B.A. , F.R.S. assisted by Alfred J. Hipktns ( of John Broadwood and Sons ) .
Received October 30 , 1884 .
Musical Scales are said to be Harmonic or Non-harmonic according as they are or are not .adapted for playing in harmony .
Most accounts of non-harmonic scales , such as the Greek , Arabic , and Persian , either ( 1 ) are derived from native theoreticians , who give the comparative lengths of the strings for the several notes , whence , on the assumption that the numbers of vibrations are inversely proportional to the lengths ( which is only approximately correct in practice ) , the intervals from note to note are inferred ; or ( 2 ) are attempts to express the effects of the intervals by the European equally tempered scale .
The former when reduced , as in Professor J. P. N. Land 's " Gamme Arabe , " 1884 , is the best that can be done without hearing the scales themselves .
The latter is utterly delusive and misleading .
Having about 100 tuning-forks , the pitch of each of which has been determined by Scheibler 's forks ( see " Proc. Roy .
Soc. , " June , 1880 , vol. 30 , p. 525 ) , and having had an opportunity of hearing the notes themselves produced on various instruments , and having had the great advantage of being assisted by Mr. A. J. Hipkins 's musical ear , which is wonderfully aeute to detect and estimate minute differences of pitch , and without which I could have done little , * I have been able , I believe for the first time , to take down the actual pitch of the notes in various existing non-harmonic scales far better than it was possible to do with the siren or the monochord , which are not only difficult to manipulate and to carry about , but at the best are very apt to mislead .
Where it was impossible actually to hear the sounds , I carefully measured the comparative vibrating lengths of the strings producing the notes on fretted instruments , whence , with by no means the same certainty , the scales could be inferred .
But I have not here noted these measurements or their results , unless I could contrast them with the intervals obtained by measuring the actual pitch of the note3 produced on the instruments themselves , as in the cases of India and Japan .
But the mere statement of the numbers of vibrations , or of the vibrating lengths of the strings producing a scale , conveys no musical notion whatever to a musician .
He wants to know how many equally * Throughout this paper , " we " and " us " relate to Mr. Hipkins and myself jointly , and all measurements of numbers of vibrations made by us rest on the judgment of Mr. Hipkins 's ear with respect to the position of the note heard between two forks , of which I had previously determined the pitch , or their Octaves .
1884.1 some existing Non-harmonic Musical Scales .
tempered Semitones , or parts of such .
Semitones , ai'e contained in the interval , so that he can realise it somewhat , as compared with the notes of a modern piano , which are intended to* be tuned in equal temperament.* This transformation is easily effected by the following brief table , premising that for brevity I use cent for the hundredth part of an equally tempered Semitone , of which there are twelve to the Octave .
To convert tabular logarithms into cents , and conversely\#151 ; Cents .
Logs .
Cents .
Logs .
Cents .
Logs .
Cent. Logs .
100 .02509 i 10 .00251 i .00025 .i .00003 200 .05017 20 .00502 2 .00050 \#166 ; 2 .00005 300 .07526 , 30 00753 3 .00075 .3 .00008 400 .10034 40 .01003 4 .00100 .4 .00010 500 .12543 50 .01254 5 .00125 *5 .00013 600 .15051 60 .01505 6 .00151 .6 .00015 700 .17560 , 70 .01756 7 .00176 .7 .00018 800 .20069 80 .02007 8 ; 00201 .8 .00020 900 .22577- 90 .02258 9 .00226 .9 .00023 1000 .25086 1100 .27594 1200 .30103 take the logarithm of the interval ratio and seek the next least in the first column of the table ; then the next least to the difference , and so on , taking the cents opposite .
Generally it suffices to take to the nearest cent , as that expresses an insensible interval .
Thus , if the numbers of vibrations are 440 and 528 , the difference of their logs , is '07918 ; the next least in the first column , '07526 , gives 300 cts .
, with remainder , '00392 ; the next least to which in the second column , -'00251 , gives 10 cts .
, and remainder , '00141 ; the next least to which in the third column , '00125 , gives 5 cts .
, and remainder , '00016 , which in the fourth column gives '6 ct .
Hence the interval is 315'6 cts .
, for which usually 316 cts .
is sufficient to write .
Now this shows that the interval contains 3 equal Semitones , and 16 hundredths of a Semi* The first person to propose the measuring of musical intervals by equal Semitones was , I believe , de Prony , but I have not been able to see his pamphlet ; the next was the late Professor de Morgan ( " Cam .
Phil. Trans. , " x , 129 ) , from whom I learned it , and I employed it in the Appendix of my translation of Helmholtz , by the advice of Mr. Bosanquet .
Having found that two places of decimals sufficed for most purposes , I was led to take the second place , or hundredth of an equal Semitone as the unit , and 1 have extensively employed this practice , here for the first time published , with the greatest advantage .
In fact , I do not know how I could have expressed the results of the present investigation in any other brief and precise , and at the same time suggestive , method .
370 Mr. A. J. Ellis .
Tonometrical Observations on [ Nov. 20 tone more .
It is therefore the just minor Third , and it is written between the notes that form it , thus : A 316 G. It is convenient for comparison with what follows , to have the following just intervals expressed in cents :\#151 ; Intervals .
The Skhisma ... ... ... ... . .
The Comma of Didymus , The Pythagorean Comma The Septimal Comma , .
Quartertone ... ... Small Semitone , f f ... . .
Pythagorean Limma , -|-|f Small Limma , --o f. .
... .
Diatonic Semitone , -j-S- ... Pythagorean Apotome . .
Great Limma , S\#163 ; ... ... .
The Trumpet f Tone , | -f The minor Second , y. .
, The major Second , f. . . .
Septimal minor Third , Pythagorean minor Third , ^ Just minor Third , -jj- ... Cents .
Intervals .
2 Just major Third , -J 22 24 Pythagorean major Third , Grave Fourth , 27 Just Fourth , % 50 Septimal Fifth , Z 70 Tritone , 90 Grave Fifth , 92 Just Fifth , ^ 112 Acute Fifth , ... . . .
... .
114 Just minor Sixth , \#166 ; S 134 Just major Sixth , 151 182 204 Pythagorean major Sixth , Natural minor Seventh , \ . .
Minor Seventh , ... ... . .
267 294 816 Just major Seventh , . .
\lt ; .
Pythag .
major Seventh , Octave .
Cents .
In each scale I give the measured number of vibrations with , occasionally , the millimetres in the vibrating lengths of string , the cents in the interval from note to note , and the sum of those cents from the lowest note to the note considered .
From the latter , considering the lowest note to be c in all cases , it is easy to deduce the name Of the nearest equally tempered note , and show how many cents must be added to it or subtracted from it to give the note heard , by remembering that\#151 ; c cjkor dl* d d\amp ; or eh e f fit or oh g o# or ah =0 100 200 300 400 500 600 700 *800 a aft or \amp ; [ , b 900 1000 .
1100 It must be borne in mind that I give the actual intervals heard from or measured on actual instruments , and that these , we may safely say , never represent the intervals intended by the tuner , within from 5 to 20 cents either way , on account of the extreme difficulty of precise tuning , especially when the intervals are non-harmonic .
European ears are at present satisfied , on our theoretical equally tempered scale , with Fifths too flat , and Fourths too sharp by 2 cts .
, .
with major Sevenths too sharp by 12 cts .
; major Thirds too sharp by 1884.1 some existing Non-harmonic Musical Scales .
371 14 cents , and major Sixths too sharp by 16 cents , while of course the minor Sixths are 14 cts .
too flat , and the minor Thirds 16 cts .
too flat .
That is to say , these would be the errors if the tuning were perfect .
The practice , as I have determined by actual measurement , is necessarily far from being restricted to these limits .
Hence the results here given have to be compared with many other results from other instruments of the same kind , tuned by different tuners before the intended intervals could be , if they ever can be , satisfactorily determined .
In the meantime we know that native ears have actually been satisfied by the intervals here given .
It must also be remembered that as the tones heard were often exceedingly brief ( as from wood harmonicons ) , or very impure , being mixed with inharmonic proper tones ( as from metal harmonicons , kettles , gongs , \amp ; c. ) , it was generally impossible to count beats , and often even exceedingly difficult to tell within what pair of forks the note heard really lay , so that there is a possible error of two vibrations occasionally , but , thanks to the acuteness of Mr. Hipkins 's ear , it is not probable that the error at any time exceeds one vibration in a second .
The number determined is therefore purposely given Only to the nearest integer .
I. Arabia and Syria .
The theoretical account of Arabic scales is admirably given in Professor Land 's " Gamme Arabe .
" It there appears that one Zalzal , more than a thousand years ago , being dissatisfied with the ordinary division of the Fourth , as\#151 ; .
0204 D 90 E\\gt ; 114 90 0 204 294 408 498 ( where the figures between give the number of cents from note to note , and the figures below give the number of cents from the lowest note ) , introduced a division , which , carried out to the Octave , amounted to\#151 ; C 204 D 151 qE 143 F204 151 q 143 B\\gt ; 204 0 , 0 204 355 498 702 853 996 1200 where q Eand q A mean about a quarter of a tone less than ( or before coming to ) E and A , and in the same way Eq , Aq would mean a quarter of a tone beyond E and A. ( In musical notes q will become \lt ; | , a turned |\gt ; .
) In later periods this was tempered to a division of the Octave into 24 equal Quartertones , as we learn from Eli Smith , an American missionary at Damascus , who translated Meshaqah 's treatise in the " Journal of the American Oriental Society , " 1849 , vol. i , pp. 171\#151 ; 217 .
The scale therefore becomes\#151 ; 872 Mr. A. J. Ellis .
Tonometrical Observations oft [ Nov. 20 0 200 D 150 qE 150 F 200 G 150 q 150 B\ , 200 G 0 200 350 500 700 850 1000 1200 although in the Middle Ages a different scale prevailed in Arabia , to which I need not further allude .
Now between Zalzal 's time and this mediaeval alteration the Crusaders brought the Syrian bagpipe to England , and after it had passed out of fashion in England , it became the national instrument of the Highlands of Scotland .
Such an instrument , made by Macdonald , of Edinburgh , and obligingly played to us by its possessor , Mr. Charles Keen , the well-known artist , yielded on examination the following results :\#151 ; Highland Bagpipe .
Vib.* ... ... .
395 441 494 537 587 662 722 790 882 From vib . . .
g ' 191 a ' 197 V 144 c " 154 d " 208 e " 150 f " 156 191 a " Sums from a'.\#151 ; 191 0 197 341 495 703 853 1009 1200 Tempered ..\#151 ; 200 0 200 350 500 700 850 1000 1200 Notes ... .
c d qe f g l\\gt ; c ' The tempered form , therefore , coincides with the Damascus form of Zalzal 's scale , which I did not discover till long afterwards .
The theory of this scale is lost , but it is usual to make g ' to a ' rather less than a whole tone , while the two drones , an Octave and two Octaves below a ' , necessitate a pure Fifth , a ' 702 e " .
Zalzal divided a Pythagorean minor Third of 294 cents into 151 and 143 cents ; the modern instrument divides the just minor Third 316 cents , probably , into 151 and 165 parts .
We thus get a possible rationalised form of the bagpipe scale , the first attempted , so far as I know .
As usual in bagpipe music , I begin the scale on a ' .
I have calculated the vibration to the same base a ' 441 vib .
, for both tempered and rational vibrations , to show how close they are to the observed :\#151 ; nationalisation of the Bagpipe Scale .
Observed vib . .
441 494 537 587 662 722 790 882 Tempered vib. .
441 495 540 587 661 721 786 882 Rational vib ... 441 496 541 595 662 722 794 882 Notes ... ... a'204 V 151 c " 165 d " 182 e " 151 f " 165 g " 182 a " Sums of cents .
0 204 355 520 702 853 1018 1200 Ratios ... ... a'8:9 V 11:12 c " 10:11 d"9:10 e " 11:12/ " 10:119:10 Ratio from a'. .
1 8:9 22:27 20:27 2:3 11:18 5:9 2 * In these tables the line " vib .
" contains the number of vibrations determined by us .
The line " from vib : " contains the notes , in this case those usually given in bagpipe music , but generally merely distinguished by Roman numerals , I , II , III , \amp ; c. , with the interval between them in cents .
The line " sums " gives the sums of these cents , interval by interval , that is , the interval between each note and the lowest .
The line " tempered " shows the nearest intervals on an equally tempered scale of 24 Quartertones in the Octave .
The " notes " sometimes added , as those due to taking 0 as c , as already explained .
1884 .
] some existing Non-harmonic Musical Scales .
378 II .
India .
There are two distinct kinds of scales in India , those of harmonicons , most probably from hill tribes , and those of the stringed instruments belonging to the conquering race .
Balafovg from Patna in the South Kensington Museum , a wooden harmonicon strung over a beautifully carved case , consisting of 25 bars ( of which we measured 14 ) containing 3 Octaves and 3 notes .
The Roman numerals II , III , \amp ; c. , indicate the successive bars , I was not measured .
Vib ... ... ... ... .
158 176 194 214 233 259 279 320 From vib ... ... . .
II 187 III 169 IV 170 V 147 VI 183 VII 129 VIII 237 IX Sums ... ... ... ... .
0 187 356 526 673 856 985 1222 Vib ... ... ... ... .
320 355 391 434 484 .
531 582 From vib ... ... IX 180 X 167 XI 181 XII 189 XIII 160 XIV 159 XV Sums , less 1222. .
0 180 347 528 717 877 1036 Observe IV 356 cents , and VII 856 cents , which compare with Zalzal in I. Arabia .
All the Octaves were too sharp .
The old Indian stringed instrument is the Vina with frets to -S- inch high , so that by pressing the string behind the fret the pitch can be greatly altered .
These frets are shiftable , but are usually fastened with wax .
I measured the vibrating lengths of string of many , but I consider the resulting scales not sufficiently trustworthy for record here .
This pressing behind the fret is constantly employed to sharpen the pitch by a quarter or half a Tone .
The modem Sitdr , which has practically superseded the Vina , is a very long-necked guitar with movable frets .
These frets are set for the rag or ragini ( tune , key , or mode ) in which the musician is going to play .
They are high enough above the fingerboard to allow pressure behind to exert a sensible effect , but the ordinary method of raising the pitch is to deflect the string by moving the finger with the string transversely along the fret .
As , however , the frets are properly set , this deflection is used only for grace notes at the end , suddenly raising the pitch about a quarter of a Tone and returning it to its former position .
H.H. Raja Ram Pal Singh was kind enough to bring his sitar ( which he left with me ) , and setting it in five different manners to play Indian airs to us .
After he had done so I measured the position of the frets , so that I could return them to their places .
Afterwards we sounded each note , took its pitch , and determined the scale by my forks .
This , I believe , is the first time that this has been done for any Indian instrument .
The pitch for the open string was not the same as that used by the Raja , for these measurements were not taken till long afterwards , but the relative pitch remained the same .
This string , which w ; as an English pianoforte steel wire , replacing the Indian steel wire which was broken , was too thick , and this interfered somewhat with the .
setting .
As I had calculated the 374 Mr. A. J. Ellis .
Tonometrical Observations on [ Nov. 20 , intervals in cents from the vibrating lengths , I add these also in millimetres to show how unsatisfactory are the results thus obtained .
The I , II , \amp ; cM number the frets used which , however , begun at about the interval of a Fifth from the open string .
First setting of the Sitar\#151 ; Vib .
lengths ... 616 554 503 452 417 380 339 316 mm. From lengths ... I 184 II 167 III 185 IV 140 V 161 VI 198 VII 122 VIII Sams ... ... ... ... 0 184 351 536 676 837 1035 1155 Vib ... ... ... ... .
393 437 479 535 584 650 731 800 From vib ... ... .
I 183 II 159 III 191 IV 152 V 186 VI 203 VII 156 VIII Sums ... ... ... ... 0 183 342 533 685 871 1074 1230 Second setting of the Sitar\#151 ; Vib .
lengths ... 616 554 530 452 417 380 355 316 mm. From lengths ... I 184 II 77 III 276 IV 140 V 161 VI 118 VII 201 VIII Sums ... ... ... ... 0 184 261 537 677 838 956 1157 Vib ... ... ... ... .
393 437 460 535 584 650 693 800 From vib ... . .
I 183 II 89 III 262 IV 152 V 186 VI 111 VII 249 VUI Sums ... ... ... . .
0 183 272 534 686 872 983 1232 Third setting of the Sitar-\#151 ; Vib lengths ... 616 577 .
518 452 417 389 352 318 mm. From lengths ... I 113 II 187 III 236 IV 140 V 120 VI 173 VII 177 VIII Sums ... ... ... ... 0 113 800 536 676 796 969 1146 Vib ... ... ... ... .
393 419 471 535 584 634 707 785 From vib ... . .
I 111 II 203 III 220 IV 152 V 142 VI 189 VII 181 VIII Sums ... ... ... ... 0 111 314 534 686 828 1017 1198 Fourth setting of the Sitar\#151 ; Vib .
lengths ... 612 552 500 466 415 368 337 318 mm. From lengths ... I 179 II 171 III 122 IV 201 V 208 VI 152 Vil 100 VIII Sums ... ... ... ... 0 179 350 472 673 881 1033 1133 Vib ... ... ... ... .
397 439 486 523 594 671 737 786 From vib ... . .
I 174 II 176 III 127 IV 220 V 211 VI 162 VII 111 VIII Sums ... ... ... ... 0 174 350 477 697 908 1070 1181 Fifth setting of the Sitar\#151 ; Vib .
lengths ... 607 574 492 461 408 384 332 312 mu ' .
From lengths ... I 97 II 267 III 113 IV 212 V 105 VI 252 VII 108 VIII Sums ... ... ... ... 0 97 364 477 689 794 1046 1154 Vib ... ... ... ... .
395 416 488 525 594 620 737 784 From vib ... . .
I 90 II 276 III 127 IV 214 V 74 VI 299 VII 107 VIII Sums ... ... ... ... 0 90 366 493 707 781 1080 1187 I would draw attention to the great difference in all cases between the two last intervals , I to VII , and I to VIII , as calculated from the lengths of the strings and the number of vibrations .
This arose from the string lying naturally further above the frets for the last notes , and hence the tension being more increased by pressing the string to the fret .
Also observe how nearly III approaches to 350 cents in the first , fourth , and fifth settings , and VI to 850 cents in the first , second , 1884 .
] some existing Non-harmonic Musical Scales .
375 and third settings , taking all from the intervals heard .
The Indian system of scales is very complex , and differs much from the European .
III .
Singapore .
Mr. Hipkins received a Balafong or wood harmonicon direct from Singapore , consisting of 24 bars forming 3 Octaves and 3 notes .
We measured the central Octave , beginning at bar 8 , as follows:\#151 ; Observed vib ... .
312 344 382 427 470 523 569 626 From vib ... ... .
I 169 II 181 III 193 IV 166 V 185 VI 146 VII 165 VIII Sums ... *..\#171 ; \#187 ; \#171 ; ..* 0 169 350 543 709 894 1040 1205 Tempered vib. .
312 340 382 429 467 525 572 624 From vib ... ... .
I 150 II 200 III 200 IV 150 V 200 VI 150 VII 150 VIII Sums ... ... ... ... .
0 150 350 550 700 900 1050 1200 The tempered form is given to show that this is one of the Quarter-tone systems , and the tempered vibrations were calculated to show how near they are to the observed .
IV .
Burmah .
The Patala or wood harmonicon of 25 small neat bars in the South Kensington Museum , No. 1630\#151 ; '72 , " Engel , " p. 16 , who gives the scale wrongly .
We began at the seventh bar from the end , and took an Octave thus:\#151 ; Vib ... ... ... . .
300 332 367 408 451 504 551 616 Bars ... ... ... . .
I 176 II 174 III 183 IV 174 V 192 VI 154 VII 193 VIII Sums ... ... ... .
0 176 350 533 707 899 1053 1246 The Octave is very sharp , and bars 15 , 16 , 20 were sharp Octaves of II , III , VII , bar 16 being very sharp indeed .
Otherwise the Octaves were fair .
A Balafong , in South Kensington Museum , with a box decorated with Burmese ornaments , 22 bars , containing 3 Octaves and 1 note .
Twelve bars measured from 4th to the 15th .
The first 5 formed the end of an Octave .
Vib ... ... ... ... ... ... 287 258 282 318 353 Prom vib .
... . .
... ... IV 147 V 154 VI 208 VII 181 VIII Sums ... ... . .
... ... ... 506 653 807 1015 1196 Vib ... ... ... .
353 377 432 485 525 573 641 705 From vib ... .
VIII 114 IX 236 X 200 XI 137 XII 151 XIII 194 XIV 164 XV Sums ... ... ... 0 114 350 550 687 838 1032 1196 The sums in the first line have been found by subtraction from that under VIII , which was assumed to be the same as that under XV .
The different construction of the corresponding parts of the Octave is thus shown .
The Keay Wine in South Kensington Museum consists of 15 kettles or gongs resembling the Javese bonangs , arranged in a circle .
Ill* was 376 Mr. A. J. Ellis .
Tonometrical Observations on [ Nov. 20 , cracked , and its pitch is doubtful , as was also that of V* .
II and III* , as the latter stood , were practically identical .
First oct.vib ... 303 333 334 377 416 449 506 602 From vib ... ... I 163 II 5 III* 210 IV 170 V* 132 VI 267 VII 301 VIII Sums ... ... ... ... 0 163 168 378 548 680 947 1248 Second oct .
vib .
602 622 648 719 796 867 990 1032 From vib ... . .
VIII 57 IX 71 X 180 XI 176 XII 148 XIII 230 XIV 72 XV Sums ... ... ... . .
0 57 128 308 484 632 862 934 The kettles were probably all out of tune .
V. Siam .
The Hanat in South Kensington Museum is a wood harmonicon with 19 bars , scale wrongly described in " Engel , " p. 316 .
Bar XIII* was of a different kind of wood , and had evidently been inserted as a substitute for the Octave of VI , but was too sharp .
First oct .
vib ... .
323 348 379 433 491 504 585 666 From vib ... ... VI 129 VII 148 VIII 231 IX 218 X 45 XI 258 XII 225 XIII* Sums ... ... ... ... . .
0 129 277 508 726 771 1029 1254 Second oct , vib .
666 748 794 From vib ... ... XIII* 201 XIV 103 XV Sums ... ... ... ... . .
0 201 304 The scale is enigmatical .
VI .
West Coast of Africa .
This is inserted out of geographical position , because it is a solitary example from Africa , and resembles those immediately preceding in character .
A Balafong in South Kensington Museum , No. 1080 , 1080a\#151 ; '68 , " Engel , " p. 154 , who describes the scale wrongly .
We measured nine bars\#151 ; .
Observed yib. .
327 357 386 445 497 547 596 654 714 From vib ... VIII 152 IX 135 X 246 XI 191 XII 166 XIII 149 XIV 161 XV 152 XVI Sums ... ... ... . .
0 152 287 533 724 890 1039 1200 1352 Tempered vib .
327 357 389 449 504 550 600 654 From vib ... VIII 150 IX 150 X 250 XI 200 XII 150 XIII 150 XIV 150 XV Sums ... ... ... . .
0 150 300 550 750 900 1050 1200 where the tempering shows that the scale belongs to the system of Quartertones .
VII .
Java .
The scales were observed from the instruments of the Javese Gramelang or band , at the Aquarium , in November , 1882 , and formed the commencement of these investigations .
We were materially assisted by work done on the same instruments ( but without determining pitch ) by Mr. W. Stephen Mitchell , M.A. , of Gonville and Cains College , Cambridge , and by determinations with the monochord of similar instruments in Holland by Professor J. P. N. Land ( who 1884 .
] some existing Non-harmonic Musical Scales .
377 also gave me much information ) , assisted by Dr. Onnes , both of Leyden .
Professor Land also kindly communicated the results of the measurements by Dr. Loman and Dr. Figee , both of Leyden .
These measurements of distinct instruments are annexed in a reduced form .
There are two entirely different Javese orchestras which cannot play together .
We examined three sets of instruments from each\#151 ; the Gambang , or wooden harmonicon , the Sdron and Slentem , or metal bar harmonicons , and the Bonang , or set of kettles\#151 ; while in Leyden a Gender ( another metal harmonicon ) and a different Sdron were examined .
The first orchestra played Salendro , the second Pelog scales , both Pentatonic ; but , as will be seen , completely different .
The first had only five notes in the Octave , the second had seven , but used only five at a time , just as Europeans have twelve , but use only seven at a time .
The first has no interval between consecutive notes so small as a major Second , or so large as a minor Third .
The second has between two consecutive notes of its seven , approximatively two Semitones , ( no Tone ) , three Three-quartertones , and two minor Thirds .
The first is very uniform , the second very diverse in its intervals .
First oi* Salendro Scales .
Obs. vib .
* Out of tune .
+ Not recorded .
Gambang ... ... ... *268 308 357 411 470 *535 Saron ( E. \amp ; H. ) . .
272 308 357 411 471 543 Sl\amp ; ntem ... ... ... ... .
270 308 357 411 469 540 Mean ... ... ... ... ... 270 308 357 411 470 540 From Mean ... ... ... . .
I 228 II 256 III 244 IY 232 Y 240 I ' Sums ... ... ... ... ... . .
0 228 484 728 960 1200 GSnder , lower oct , .
- I 191 II 251 III 249 IY 261 Y 220 V Gfinder , upper oct . .
I 219 II 256 III 261 IY 223 Y 288 I ' Saron ( Land ) ... ... . .
I 270 II 200 III 266 IY 239 Y 243 Saron ( Figee ) ... .
I 275 JI 210 Illf IVf Y 243 I ' Tempered vib ... . .
270 310 356 409 470 540 From vib .
... ... ... . .
I 240 II 240 III 240 IV 240 V 240 I ' Sums ... ... ... ... ... . .
0 240 480 720 960 1200 This tempered form seems to have been that aimed at .
It is easily tuned when the ear has become accustomed to the flat Fourth of 480 cents .
Tune up I 480 III , and III 480 V. Then from the Octave V tune down I'\#151 ; 480 IV , and IV\#151 ; 480 II .
Observe that the Fourth is flat and the Fifth sharp , and that V is nearly the natural harmonic Seventh of 969 cents .
These are also points of distinction from the next set .
378 Mr. A. J. Ellis .
Tonometrical Observations on [ Nov. 20 , Second or Pelog Scales .
Obs. Vibrations .
* Out of tune .
Gambang ... ... ... .
*283 *311 365 391 416 448 *532 *566 Bonang ... ... ... . .
278 302 861 390 417 448 526 556 Saron ... ... ... ... 279 302 360 387 414 447 524 558 Adopted ... ... ... . .
279 302 361 389 415 448 526 558 From adopted ... ... I 137 II 309 III 129 IV 112 V 133 VI 278 VII 102 V Sums ... ... ... ... .
0 137 446 575 687 820 1098 1200 .
Scales .
P51og ... ... ... . .
I 446 Ill 129 IV 112 V 411 ... VII 102 V Dantsoe ... ... ... I 137 II 550 ... ... ... ... V 133 VI 278 VII 102 V Bern ( E. \amp ; H.). .
I 137 II 438 ... IV 112 V 411 ... VII 102 I " ( Loman ) ... .
I 147 II 416 ... IV 96 V 429 ... VII 112 I ' Barang(E .
\amp ; d H. ) ... I 137 II 438 ... IV 112 V 133 VI 380 ... * I ' " ( Loman ) ... I 151 II 426 ... IV 111 V 179 VI 333 ... I ' Miring ... ... ... .
I 446 ... Ill 129 IV 245 ... VI 278 VII 102 I ' Menjoera ... ... ... I 137 II 309 III 129 IV 523 ... ... ... ... .
VII 102 V Tempered vib ... .
279 304 362 395 418 443 527 558 From vib ... ... . .
I 150 II 300 III 150 IV 100 V 100 VI 300 VII 100 1 ' Sums ... ... ... ... 0 150 450 600 700 800 1100 1200 After giving the three sets of vibrations observed I give that adopted , which is the mean of the second and third set , as the Grambang was evidently rather out of tune , and then the scale of all the seven notes answering to the chromatic scale of our pianos .
Then follow the names of the scales really used , formed by selecting five notes from these .
Pelog and Dantsoe ( pronounce Dutch oe as our oe in shoe ' ) are given only from our own observations .
In Bern and Barang , Dr. Loman 's observations made with the monochord in 1879 on another set of instruments are added in a reduced form .
These four scales are certain .
Miring and Menjoera ( pronounce Dutch joe like the English word ) are conjectural restorations from imperfect indications communicated to me by Professor Land .
Finally , I have added a rather hazardous tempering , and shown by calculating the vibrations from it , that it does not materially misrepresent the observed .
In these scales the Fourth , IV 575 cents , is nearly the tempered Tritone 600 cents , and the Fifth , V 687 cents , is flatter even than the tempered Fifth 700 cents .
This is exactly contrary to the Salendro scale .
Vet I observed one of the players selecting the right bar for his scale by holding it up and tapping it with his finger , showing that the pitch was quite familiar to him .
VIII .
China .
Without entering upon any discussion on the very vexed question of Chinese music , I confine myself to giving the scales which ( by the kind permission of Mr. J. D. Campbell , one of the Commissioners of Chinese customs representing China at the International Health Exhibition this year , and with the assistance of the secretary , Mr. Neumann ) , we were able to have played to us by the Chinese 1884.1 some existing Non-harmonic Musical Scales .
379 musicians attached to that court , in July and August , 1884 , at four specially arranged meetings , on their own instruments , together with observations on a duplicate of one of them at the South Kensington Museum , and a set of bells belonging to Mr. Hermann Smith .
1 .
Transverse Flute or Ti-tsu , with seven finger holes and an embouchure , open at both ends .
Probably in actual playing some of the notes may have been varied by half or quarter covering of the finger-holes .
The Heptatonic scale played is given first , and then the notes selected for the more usual Pentatonic scale .
Vib .
... ... ... ... 240 266 292 311 352 401 454 479 From vib .
... ... I 178 II 161 III 109 IV 214 V 226 VI 215 VII 93 I ' Sums ... ... ... ... . .
0 178 339 448 662 888 1103 1196 Pentatonic_____ I 178 II 270 ... IV 214 V 226 VI 308 ... I ' 2 .
Oboe or So-na , played with a short reed , having seven finger-holes in front and two thumb-holes behind , a loose brass cone of considerable size covered the lower end .
Said to be a modern instrument .
Sound and intervals resembling the bagpipes .
Vih ... ... ... ... 400 435 475 516 578 640 719 808 From Tib .
______ I 145 II 152 III 143 IV 197 V 176 VI 201 VII 202 I ' Sums ... ... ... . .
0 145 297 440 637 813 1014 1216 Tempered vib ... 400 436 476 519 582 635 713 800 From vib ... . .
I 150 II 150 III 150 IV 200 V 150 VI 200 VII 200 I ' Sums ___________ O 150 300 450 650 800 1000 1200 On this instrument as thus played there was nothing approaching a Fourth of 498 cents , or a Fifth of 702 cents .
It must have been modified in playing to work with the flute .
Both were orchestral instruments .
3 .
Reed Mouth Organs or Sheng ( rhymes to , and often so called ) , a gourd with its top cut off , and covered with a flat board , in which were inserted 13 pipes , 11 of which had free reeds , which sounded on blowing ( or sucking ) through the mouthrhole , and stopping a hole in the pipe which the player intended to sound .
The lengths of the pipes are ornamental , an internal slot determining the real lengths .
The two " dummies " were for holding .
First oct .
vib ... .
450 508 547 600 680 760 820 899 From vib ... ... .
I 210 II 128 III 160 IV 217 V 193 VI 132 VII 159 I ' Sums ... ... ... ... .
0 216 338 498 715 908 1040 1199 Second oct .
vib .
899 ' 1017 1110 1232 From Tib ... ... .
I ' 214 II ' 151 IIP 182 IV ' Sums ... ... ... ... ... 0 214 365 547 Tempered vib ... 450 505 551 601 674 757 825 900 From vib ... ... .
I 200 II 150 III 150 IV 200 V 200 VI 150 VII 150 I ' Sums ... ... ... ... . .
0 200 350 500 700 900 1050 1200 Here we have a perfect Fourth , ' IV 498 cents , and a good but sharp Fifth , V 715 cents .
But the instrument , if in tune ( small free reeds easily fall out of tune ) , belonged to the Quartertone system .
VOL. xxxvii .
2 c 380 Mr. A. J. Ellis .
Tonometrical Observations on [ Nov. 20 , 4 .
First Chime of Small Gongs or Yan-lo , a set of 10 small gongs about the size and shape of cheese-plates , arranged with I at the top , II , III , IY in the first row , from left to right behind , where they were struck with a wooden hammer , and then Y , YI , YII in the second , and YIII , IX , X in the third row , all hung in a square wooden frame .
The Chinese musician played in the order of pitch , omitting IX and I. Vib ... ... ... ... 449 495 555 568 630 663 703 712 830 902 Fromvib ... ... .
VIII 169 V 198 II 40 IX 179 IV 88 VI 101 X 22 I 265 VII 144 in Sums ... ... ... ; . .
0 169 367 407 .
586 674 775 797 1062 1208 Played ... ... ... 0 169 367 ... 586 674 795 ... 1062 1208 Here again there is no approach to a Fourth of 498 cents , or a Fifth of 702 cents .
5 .
Second Chime of Small Gongs or Yan-lo , in the S. K. Mus. , " Engel , " p. 193 , who describes the scale wrongly .
Although the instrument is of the same appearance as the last , the scale was entirely different , and the compass did not reach 750 cents .
We seemed to make out three possible scales which are annexed , but we have no means of knowing if they were designed .
One extends to a sharp and another to a flat Fifth , whilst the third reaches an exact Fourth .
The gongs are numbered as in No. 4 .
Vib .
... ... . .
794 818 912 926 1011 1022 1114 1116 1198 1216 Fromvib ... ... I 52 II 188 III 26 IV 152 VI 19 VIII 149 V 3 IX 123 X 26 VII Sums ... ... ... . .
0 52 240 266 418 437 586 589 712 738 Possible scales To sharp Fifth I 240 ... Ill J78 ... VI 171 ... ... " IX 123 X Sums ... ... ... . .
0 240 418 589 712 To flat Fifth ... ... . .
II 188 III 178 ... VI 168 ... V 152 ... ... .
VII Sums ... ... ... 0 188 366 534 686 To Fourth ... ... ... ... ... .
Ill 197 ... ... ... .
VIII 152 ... IX 149 ... VII Sums ... ..*. .
0 # 197 349 498 Tempered ... . .
0 200 350 500 The last is therefore like the first tetrachord in the bagpipe scale , dividing the Fourth into a Tone and two Three-quartertones .
There are , however , several curious intervals .
YI 19 YIII nearly a comma of 22 cents .
Ill 26 IY nearly of a major Tone of 204 cents .
I 52 II exactly J of a major Tone of 204 cents .
II188 III and III 178 IY are both nearly the minor Tone of 182 cents .
I 240 III is an exact pentatone , or i Octave , as in the tempered Javese Salendro scale .
II 385 YIII is an excellent major Third of 386 cents .
I 586 Y and I 589 IX are both nearly the Zald of 588 cents , on the second string of the Arabic lute .
1884 .
] some existing Non-harmonic Musical Scales .
381 I 738 VII , the complete compass , is exactly the 49th harmonic reduced to the same Octave , which is of course only a curious coincidence .
6 .
Dulcimer or Yang-chin , exactly like the ordinary dulcimer ( see figure in Grove 's " Dictionary of Music , " i , 469 ) , with four wires to each note forming two Octaves , the longer wires passing under the bridge which limits the shorter .
It is struck with elastic hammers .
The instrument being out of tune was tuned for us by the musician who played No. 7 , according to the Chinese names of the scale in Dr. William 's Middle Kingdom , which are there interpreted as the major scale of F^ .
If the conjectural just scale be correct , this would be the scale of B\\gt ; major , beginning on its second note 0 , and is therefore comparable to the Japanese R'itsusen , which is the scale of G major begun on its second D.* Chinese names ... He sz ' f chang ch6 kung fan liu .
Vib 205 226 240 272 300 340 364 409 From vib I 169 II 105 III 217 IV m V 217 VI 118 VII 202 I* Sums .
0 169 274 491 661 878 996 1198 Conjectured Just Vib 205 228 243 273 304 342 364 410 From vib C 182 A 112 e\\gt ; 204 F 182 204 a 112 204 c Sums 0 182 294 498 680 884 996 1200 Pentatonic form .
e 182 .
A 316 F 182 204 A 316 ... c The tuner had great difficulty in tuning the semitones II 105 III and VI 118 VII , that is , in tuning the notes III and VII .
He accomplished the second more easily than the first .
The Pentatonic form consists of two disjunct tetrachords , OF , , each divided into a Tone and a minor Third .
8 .
Tamboura or Sien-tsuy a three-stringed guitar with , circular body and long neck without frets .
The strings were tuned to 239 , 266 , and 400 vib .
, making the intervals 185 and 706 cents , meant for 182 the minor tone , between the first and second , and for 702 , a Fifth , between the Second and Third , very fairly tuned indeed .
The strings were plucked with bone plectrums , attached to the first joint of thumb and forefinger , and projecting like claws .
The tone was good and very like a banjo .
Only the following pentatonic scale was played to us:\#151 ; Vib ... ... ... ... ... 320 357 400 480 536 642 From vib .
... ... . .
I 189 II 197 III 316 IV 191 V 312 V Sums ... ... ... ... .
0 189 386 702 893 1200 * In writing tones in Pythagorean intonation formed by a succession of just Fifths or Fourths from C , the ordinary letters are kept unchanged ; but for just intonation it is necessary to have a series a comma lower .
These have a subscript 1 , as Z\gt ; i , so that , in vibrations , JD^ : D\#151 ; 80 : 81 .
Similarly another series would be a comma sharper , and be written with a superior 1 , as US'k , so that , in vibrations , E\\gt ; ; E^h \#151 ; 80 : 81 .
382 Mr. A. J. Ellis .
Tonometrical Observations on [ Nov. 20 Conjectural Just Vib .
... .\#171 ; ... ... . .
320 356 400 .
480 535 640 From yib ... ... ... . .
C 182 JDl 204 JE\ 316 G 182 A1 316 c Sums ... ... ... ... ... 0 182 386 702 884 1200 Transformed sums . .
498 680 884 0 182 498 This was again so nearly just that I have conjectured a just restoration , O Dx El Q Axc : and if this is transformed , by beginning it with G , or by deducting 702 cents from each of the last sums ( previously adding 1200 cents where needed ) , we obtain the scale G 182 J. 316 G 182 D1 204 El 316 G , in which the intervals are precisely the same as in No. 7 .
9 .
Balloon Guitar or P't-p'a.\#151 ; The body of the guitar was oval .
There were four strings , the lowest tuned to 234 vib .
, and then its Fourth , its Fifth , and its Octave , but we did not test the accuracy of these intervals , which were tuned by the same musician who tuned Nos. 7 and 8 .
Near the nut were four large , round-backed , semielliptical frets , joining each other at bottom .
These the player did not use .
But on two examples .
of the S. K. Museum , I conjectured by measuring the strings , that they were intended to give such a tetrachord as\#151 ; C204 D90 E\\gt ; 114 H 90 F 0 204 294 408 498 or their just or tempered forms .
There were 12 frets on the body of the instrument .
They were high but broad"at the top .
We did not test each , but merely took down the following pentatonic scale :\#151 ; Observed vib. .
320 348 392 465 530 638 From vib ... .
I 145 II 206 III 296 IV 227 V 321 VI Sums ... ... . .
0 145 351 647 874 1195 Tempered vib. .
320 349 392 466 538 640 From vib ... .
I 150 II 200 III 300 IV 250 V 300 VI Sums ... ... ... ... O 150 350 650 900 1200 The tempered scale agrees well in all notes but V. The scale is so remarkable in every way , though it did not sound amiss , that I suspect the frets to have been inaccurately placed ; they were bits of wood roughly glued on .
This completes our observations with the Chinese musicians .
I measured also the vibrating lengths of strings in two other P'i-p'as , and also two Moon Guitars or Yin the S. K. Museum .
One of the latter seemed intended for equal temperament of 12 Semitones , and it is the only Chinese instrument which has suggested this to me ; the other looked like an attempt to divide the Octave into eight Three-quartertones , and had at any rate eight tones 1884 .
] some existing Non-harmonic Musical .
to the Octave forming nearly those intervals .
But as I did not try these with forks I do not record them .
10 .
Small Chime of Bells , belonging to Mr. Hermann Smith .
Four small bells of which the largest was 45 mm. in diameter and 13 mm. in height , arranged on a stem passing through them and framed in a lyre-shaped wire .
Yib ... ... ... . .
761 912 1004 1156 From vib ... . .
I 313 II 167 III 244 IV Sums ... ... . .
0 313 480 724 The I 313 II is nearly a perfect minor Third of 316 cents .
The III and IV give almost precisely the Javese Salendro observed III 484 , and IV 728 , so that the interval between them , 244 ceuts , is almost precisely a Pentatone of 240 cents , or Octave .
If indeed II were flatter , the notes of the bells might pass as part of such a scale .
IX .
Japan .
In the Educational Section of the International Health Exhibition of 1884 there was a considerable collection of Japanese instruments , but there were no players .
The only instruments which we could try therefore were a Sho ( the Chinese sheng ( see China , 3 ) , but different in the number and pitch and intervals of the notes ) and a Biwa , or four-string fretted lute .
The Sho we found to be out of tune , as referred to the scale exhibited , and to be impossible to blow satisfactorily .
The Biwa I first tried by measuring the lengths of the strings , and afterwards with Mr. Hipkins , by tuning the strings arbitrarily and taking the pitch from each fret .
These results I record , because in addition to the examples from India , they show very well that measurements of lengths are only an approximation to the speaking values of the strings , and that the latter vary considerably with the thickness of the strings .
This has an important bearing upon the theoretical determination of scales given by the divisions of the string .
The results for India were valuable in this respect , but they were not altogether satisfactory , because the string was English and too thick .
In the present case we had the genuine Japanese strings .
The Biwa is a large and heavy but handsome instrument , well made and finished , and answers exactly to A1 Farabi 's lutein Professor Land 's " Gamme A.rabe , " the four strings nearly coinciding at the nut , passing over a semi-circular depression to the large tuning pegs , and spreading out to a convenient distance apart by the bridge , so that the plectrum , made of hard wood , spread out like the head of a halbert , could easily be inserted between the strings , or pass over them in rapid succession for arpeggio chords for which the instrument 384 Mr. A. J. Ellis .
Tonometrical Observations on [ Nov. 20 seems to be much used in accompaniments , judging from some music written for it in Japan , on the European staff , the original of which I saw .
The diameters of the strings , which seemed to be of hard-corded silk , taken by one of Elliott 's micrometer gauges , were 1/ 65 P37 , TOG , and 0'88 mm. in diameter respectively .
The variations of interval , however , with the thickness of the string appear not to follow any precise law .
The frets were high and about 5 mm. wide of the top , made of hard wood .
I was very careful to press on the top of the fret , so that the tension of the string might not be increased , and the action should take place from the edge of the fret nearest the bridge .
But possibly I may not always have pressed near enough to the edge , so that the string was slightly lengthened and the pitch flattened .
Of course nothing like such accuracy would be reached by the player .
Lengths. .
, 843 750 709 673 637 mm. From lengths I 202 II 97 III 90 IY 95 Y Sums . .
0 202 299 389 484 Lowest string .
Yib 166 189 201 211 223 From vib * I 225 II 107 III 84 IY 96 Y Sums ; 0 225 332 416 612 Second lowest siring .
Yib 167 190 203 214 223 From vib ... ... ... ... .
I 223 II 115 III 91 IY 71 Y Sums 0 223 338 429 500 Second highest string .
Yib .
226 253 272 286 301 From vib * . . . .
I 195 II 125 III 87 IV 89 Y Sums 0 195 320 407 496 Highest string .
Yib 300 339 361 381 401 From yib I 212 ri 109 III 93 IY 89 Y Sums 0 212 321 414 503 Mean from yib I 214 II 114 III 89 IY 86 Y Sums of mean 0 214 828 417 503 Possibly I 204 II 114 III 90 IY 90 Y Sums 0 204 318 408 498 Hence the division was probably meant for Pythagorean , the last sums giving C D DS E F , which should have been C D E\\gt ; E F , that is , the second Semitone should have been of 114 cents , and the first of 90 cents .
Now it appears from the Report of Mr. Isawa , Director of the Institute of Music , Tokio , Japan ( founded October , 1878 ) , an English translation of which , prepared at the Institute , was in the Section , that Japanese theory considers its Semitones to be 12 equal 1884 .
] some existing Non-harmonic .
Musical Scales .
885 divisions of the Octave , just as in Europe we so consider our 12 Semitones.* Hence these divisions are taken , and are used as\#151 ; C 200 D 100 E\\gt ; 100 E 100 0 200 300 400 500 as they would he played on the pianoforte .
This Report contains an account of the Japanese scale , from which , to complete this notice of Japan , although not tonometrically observed , 1 may cite the following , where all notes may be provisionally considered as those on the piano .
Classical Scales .
Riosen ... ... ... ... ... ... ... E E C$ d In descending often ... ... ... E E Pentatonic ... ... ... ... ... .
E E A B d Ritsusen ... ... ... ... ... ... . .
E E F A B C d Pentatonic ... ... ... ... ... . .
E E G A B d Popular Seals \#151 ; Heptatonic .
First Heptatonic ... ... ... . .
E E\\gt ; F G A B\\gt ; C d Second " ... ... ... ... ... E\\gt ; F G c d Popular Scales\#151 ; Pentatonic .
Hiradioshi ... ... ... ... ... .
G A B\\gt ; E G Akebono I ... ... ... ... ... ... G A B\ , E E G Akebono II .
... ... ... ... A JE E F A Kumoil ... ... ... ... ... ... . .
G A\\gt ; C E E\\gt ; G Han-Kumoi ... ... ... ... ... .
G A C E E'p G Iwato ... ... ... ... ... ... .
G E\\gt ; F G Han-Iwato ... ... ... ... ... G C E F G where observe the numerous examples of the most ancient Greek tctrachord of Olympos , consisting of a Semitone followed by a major Third .
* Professor Ayrton , F.R.S. , who was present when this paper was read , and who had returned from Japan only a few years ago , made some remarks to which with his permission I will here refer .
He said that it was a mistake to suppose the Japanese musical intervals to be like the European .
He had examined Japanese instruments when tuned in their different ways by natives , and taken the pitches of the notes by means of a siren , and he had found the intervals very different .
My paper in this part merely professes to give Mr. Isawa 's theory , without citing his confirmatory experiments , which I did not consider conclusive.\#151 ; A. J. E.

