SICT. XVIII. A SPEAKING MACHINE. 147 



Several attempts have been made to imitate the artic- 

 ulation of the letters of the alphabet. About the year 

 1779, MM. Kratzenstein of St. Petersburgh, and Kem- 

 pelen of Vienna, constructed instruments which articu- 

 lated many letters, words, and even sentences. Mr. 

 Willis of Cambridge has recently adapted cylindrical 

 tubes to a reed, whose length can be varied at pleasure 

 by sliding joints. Upon drawing out a tube while a col- 

 umn of air from the bellows of ah organ is passing 

 through it, the vowels are pronounced in the order, 2, 6, 

 a, o, u. On extending the tube they are repeated after 

 a certain interval, in the inverted order, u, o y a, c, i. Af- 

 ter another interval they are flgain obtained in the direct 

 order, and so on. When the pitch of the reed is very 

 high, it is impossible to sound some of the vowels, which 

 is in perfect correspondence with the human voice, fe- 

 male singers being unable to pronounce u and o in their 

 high notes. From the singular discoveries of M. Savart 

 on the nature of the human voice, and the investiga- 

 tions of Mr. Willis on the mechanism of the larynx, 

 it may be presumed that ultimately the utterance- or 

 pronunciation of mod ern*langu ages will be conveyed, 

 not only to the eye but also to the ear of posterity. 

 Had the ancients possessed the means of transmitting 

 such definite sounds, the civilized world would ^till have 

 responded in sympathetic notes at the distance of many 

 ages. 



SECTION XVIII. 



Refraction Astronomical Refraction and its Laws Formation of Tables of 

 Refraction Terrestrial Refraction Its Quantity Instances of Extraor- 

 dinary Refraction Reflection Instances of Extraordinary Reflection 

 Loss of Light by the Absorbing Power of the Atmosphere Apparent 

 Magnitude of Sun and Moon in the Horizon. 



NOT only everything we hear but all we see is through 

 the medium of the atmosphere. Without some knowl- 

 edge of its action upon light, it would be impossible to 

 ascertain the position of the heavenly bodies, or even to 

 determine the exact place of very distant objects upon 

 the surface of the earth ; for in consequence of the re- 



