712 Transactions of the American Institute. 



ous blue flame* At the least noise this flame roars, at the same time 

 flattening itself upon the gauze, and becoming almost invisible ; it is 

 very active in its responses, and being rather noisy, appeals to the ear 

 as well as the eye. It does not appear to distinguish the vowel sounds 

 so well as the vowel flame proper. To A it is extremely sensitive, 

 less so to I, but slightly sensitive to E and IT ; and O has no influence 

 upon it whatever. To the sounds coming from a small music-box it 

 dancfes in the most perfect manner, and is highly sensitive to most of 

 the sonorous vibrations which affect other flames." 



Prof. R. H. Thurston — Some interesting experiments have lately 

 been tried by Dr. Mayer at the Stevens Institute. Several methods 

 were employed for making the tubes, but the most successful one was 

 to draw out one end of a glass tube, one-fourth of an inch in diame- 

 ter, to a point, and file it, and try it, until it produced its maximum 

 effect. The most successful one was a tube of 3-16ths of an inch 

 internal diameter, tapered down nearly to one-eighth of an inch, and 

 the flame was more than a foot high. The jingle of a key, or sound- 

 ing the letter S across the room, would bring it down to about one- 

 third, and it would then make the loud rustling sound that gas flames 

 produce when the pressure is too great. He obtained uniform success 

 with brass tubes one-fourth of an inch in diameter, by soldering a 

 plate upon the top and reaming it out until the hole was the most 

 successful size. 



The President — These flames may lead to the discovery of one 

 thing hitherto entirely unexplained, with regard to the production of 

 sounds. We know that the pitch of musical sound depends entirely 

 upon the rapidity of the vibrations of the air, the " middle C " pro- 

 ducing, according to the French standard, 522 vibrations in a second. 

 But consonant sounds do not depend at all upon pitch ; and we do 

 not know on what the expression of a consonant depends. The 

 sound of S affects this flame very much, and this may help to solve 

 the problem. The only theory I can conceive of is, that the conso- 

 nant sounds produce a roughening of the waves, like the crest of waves 

 breaking upon the shore. Prof. Helmholtz has recently shown that 

 the timbre or quality of sound is determined harmonics or overtones. 



Prof. R. H. Thurston — We found that when the flame was most 

 sensitive, a variety of sounds, at different pitches, would produce the 

 same effect. It may be that greater care will make the flame still 

 more sensitive, and diminish the range, so that each tube will be 

 affected by a particular kind of sound. The subject has not yet been 

 sufficiently investigated to determine this. Within the limits of the 



