546 



SCIENCE. 



[N. S. Vol. III. No. 67. 



witli a gas cock at its neck) by immersing 

 it in water, and then filling it with illumi- 

 nating gas from the city supply. The 

 movement of the bell jar as it descended 

 into the water, and thus forced the gas to 

 the flame, was carefully guided and the 

 weight of the glass jar itself exerted a suffi- 

 cient pressure. The apparatus is extremely 

 sensitive and must be kept free from vibra- 

 tions and draughts of air. 



The use of the apparatus in the experi- 

 ments for which it was designed is to de- 

 termine the minimvim change in the ampli- 

 tude, the nature of which, i. e., whether an 

 increase or decrease of intensity can be de- 

 tected. A sound of an agreeable intensity 

 (and determined by a constant height of 

 the flame) is taken as a starting point, and 

 the subject informed that this sound will 

 very gradually increase or decrease in loud- 

 ness; he listens carefully with his head in a 

 fixed position and answers as soon as he is 

 confident of the direction of the change. 

 The operator slowly moves the index in one 

 direction or the other, takes the position 

 when the answer is given and also the time 

 of the experiment. 



How far this apparatus will be serviceable 

 for other methods of studying the sensibility 

 to sound intensities is in some measure still 

 to be determined. It may be noted, how- 

 ever, that it lends itself readily to determin- 

 ing absolute sensitiveness to sound ; for one 

 has only to note the minimum height of 

 flame giving rise to a just audible sound 

 with the head at a fixed distance from the 

 apparatus. For the method of just observa- 

 ble difference one may have the flame sound, 

 stop it, and sound it again with a slightly 

 modified intensity until the difference be- 

 tween the two sounds becomes perceptible. 

 For the method of right and wrong cases 

 the same mode of use is available, except 

 that the difference between the two sounds 

 in any one series of experiments remains 

 constant. By the method of the average 



error one should have two singing flames 

 sounding alternately, the subject attempt- 

 ing to set one of them so that the sound it 

 emits equals in intensity the standard 

 sound. To all these applications there are 

 as yet two objections : First, the sound does 

 not begin immediately after the flame is al- 

 lowed to play, but takes a considerable time 

 to rise to its full intensity. The sound 

 may be stopped instantly by suddenly 

 lowering the flame, or placing a card at the 

 top of the glass tube; but its inertia in 

 starting introduces a disturbing factor. 

 The second objection refers to the difficulty 

 of constructing two such pieces of apparatus 

 exactly alike, so that two flames vibrating 

 with the same amplitude may be regarded 

 as giving out sounds of equal intensity. 

 Neither of these difficulties is insurmounta- 

 ble, and it is to be hoped that they will be 

 solved as occasion demands. 



In concluding, it may be well to indicate 

 again that the success of the apparatus is 

 due to the fact that the change in amplitude, 

 and hence in intensity, can be directly ob- 

 served; secondly, that the sound is fairly 

 pure, of a definite pitch, agreeable and 

 continuous; and thirdly, that it may be most 

 delicately changed. All these advantages 

 result from the use of the singing flame as 

 a source of sound. 



Joseph Jasteow. 



Univeesity of Wisconsin. 



HOW NATURE REGULATES TBE RAINS. 



When American enterprise invaded with 

 its iron cavalry the mountain regions of the 

 West, many established theories were put 

 to new tests and not all sustained them- 

 selves. The relations of plant life to water 

 supply as found on the eastern half of the 

 Continent had led our fathers to believe 

 that the destruction of forests would in- 

 variably and inevitably result in the deple- 

 tion of adjacent streams and to all conse- 

 quent evils. So potent is the thick shade 



