276 PROCEEDINGS OF THE AMERICAN ACADEMY 



Several series of experiments were also tried in which a number 

 of vowel sounds in succession were impressed upon the cylinder, 

 and subsequently reproduced at different speeds, — a method 

 which enables one to compare more readily the changes in the 

 various vowels with one another. The results of these compari- 

 sons are found in Table II. 



A careful comparison of the various changes indicated in the 

 preceding tables will show a close accordance between the results 

 of different experiments. But there will be noticed a few apparent 

 discrepancies. Thus in (15) 5 rises to ou% and in (16) to a ; in 

 (15) o rises to a, and in (16) to o, while a rises to e in both series. 

 These and such other like results as exist are due chiefly to the fact 

 that it was difficult to measure the speed with any great exactness 

 at the higher rates, or to maintain this speed absolutely uniform, 

 and a slight increase in speed would suffice to change the results 

 as stated in (15) to those of (16). The vowel a sustains the same 

 change in each, notwithstanding this difference in speed, because 

 the e into which it changes is more persistent than the other vowels 

 in the series, as will be more fully explained a little later. It will 

 also be instructive to compare (16) with (17). The a in the former 

 rises to e on passing from one to three revolutions, while in the 

 latter it rises only to a#, which sound would have passed into e 

 on a slight increase of speed. In (16) the speed was doubtless a 

 trifle higher than in (17) as well as in (15). 



It must also be remembered that at certain stages a sound will 

 be on the point of passing from one recognized vowel into another, 

 so that it may be difficult to denote its sound exactly by any of the 

 conventional signs usually employed. For this reason it might be 

 preferable to substitute the symbols used by Mr. Melville Bell in 

 his "Visible Speech." 



Furthermore, we have noticed in some cases that a very trifling 

 difference in the quality of the vowel impressed upon the cylinder at 

 a low speed may cause a decided change in the vowel sound given 

 out at a considerably higher speed. Thus, the vowel e reproduced 

 at increased speeds passes into a high I and then into i. But if a 

 series of e's be spoken into the instrument, it is possible to find a 

 speed of reproduction such that some of them are heard as i 's and 

 some as *'s. On raising the speed the Vs tend to rise to i, and on 

 lowering it the t's tend to fall to 7. 



Some interesting peculiarities of different vowels were observed 

 in the course of our experiments. 



