March 15, igod] 



NATURE 



467 



fail to reproduce this error. But in point of fact this 

 phosphatic porcelain, called by the French Porcelaine 

 iendre naturelle ou Anglaise, dates back to the year 

 1748, and was made largely at Bow, and at other English 

 china factories long before the time of Spode. Numerous 

 chemical analyses of authentic specimens have proved 

 this point without the shadow of a doubt. The volume 

 would have gained greatly in scientific interest had the 

 authors introduced plates representing the microscopic 

 structure of the chief porcelains and wares One such 

 plate only is given, and that is poor. 



LETTERS TO THE EDITOR. 



{ The Editor does not hold himself responsible for opinions ex- 



. pressed by his correspondents. Neither can he undertake 



to return, or to correspond with the writers of, rejected 



manuscripts intended for this or any other part of Naturb. 



No notice is taken of anonymous communications. 1 



The Acoustic Analysis of the Vowels from the Phono- 

 graphic Record. 



Many attempts have been made to determine vowel timbre 

 from the phonographic record, with more or less of success. 

 The difficulties of transcription on a sufficiently large scale has 

 proved considerable. With the aid of Dr. F. C. Van Dyck, 



Fig. I. — A, section of phonograph wax cylinder; b, surface of the wax ; 

 c, boltom of .he furrow, with undulations much exaggerated; 

 D, sapphi.e koob of tracer ; e, rigid lever bearing adjustable mirror ; 

 F, adjustable plane mirror ; <;, source of light ; h, plate with pin-hole 

 through which light passes ; j, convex lens with conjugate foci at h 

 and L ; k, revolving drum carrying bromide paper ; L, surface of paper 

 on which point of light leaves the sinuous trace. 



Professor of Physics, in Rutgers College, I have succeeded in 

 constructing an apparatus at once simple and satisfactory, and 

 have reached results that seem worthy of consideration. 



The vowel curves obtained are large enough to be measured 

 with precision, and the method employed allows greatly in- 

 creased enlargement if desired. The apparatus I have used is 

 illustrated diagrammatically in the accompanying figure (Fig. i). 

 The most essential part is simply the automatic reproducer of 

 the Edison Phonograph from which the diaphragm is removed, 

 and upon the tracing lever of which a rigid arm is fastened 

 vkhich bears an adjustable mirror. If now the sapphire knob 

 of the tracer is made to follow a record z>ery slowly, the mirror 

 will be deflected back and forth in correspondence with the 

 undulations of the record furrow, and by means of a narrow 

 beam of light reflected from it, and focussed upon a strip of 

 moving bromide paper, a sinuous photographic frace is ob- 

 tained exactly corresponding to the bottom of the furrow in the 

 waK cylinder. 



It is manifest that this method of transcription makes pos- 

 sible a very great enlargement. In the reproducer the short 

 arm of the lever, i.€. the difference between the sapphire trac- 

 ing knob and the fulcrum, is about one-eighth of an inch in 



NO. 1585, VOL, 61] 



length. Working with a beam of light of the length of ten 

 feet gives therefore an enlargement of about one thousand 

 times. By making the short arm one-sixteenth of an inch and 

 the working distance thirty feet we may without great difficulty 

 multiply ordinates by six thousand. As yet I have not found it 

 necessary to enlarge more than one thousand times. 



Since the actual depths of the hollows made by the phono- 

 graph recorder in the wax for a good vowel tone are often as 

 much as one two-thousandth of an inch, and are generally 

 greater than one twenty-thousandth of an inch, an enlargement 

 of one thousand times gives for such tones a curve of which the 

 ordinates are easily measured with precision. 



It is not possible in the space at my command here to describe 

 the details of my apparatus or of my experiments. Those who 

 may be interested I refer to a more extended account which 



^J\PJ\f'J\Af\Al\J\^ f\i\/\f\JV\AAAf\ 510 



will shortly appear in the German Phonetic Journal, Die 

 Neueren Sprachen, and to a short description of apparatus and 

 results in the Physical Review. 



I have obtained many records of all the principal vowels of 

 English speech, but have not yet completed work on any 

 vowel save a (as in father) far enough to make even a prelimin- 

 ary report. The accompanying diagram gives a few specimen 

 a-curves (Fig. 2). The pitch is indicated by the vibration 

 number at the side of each wave-section. The quality of the 

 vowel represented is the same in each case, as far as possible, at 

 different pitches, and by different voices. The record in each 

 case gave in phonographic reproduction a clear and unmistakable 

 a (as in father). The sections shown represent about the one 

 fifty-sixth part of a second, and the whole sheet has been re- 

 duced for convenience to about one-quarter of the actual size. 

 The traces which I obtained and measured for analysis, by 



