606 



SCIENCE 



[N. S. Vol. XL. No. 1034 



of species here is in direct proportion to the 

 amount of rainfall. The stream gradient at 

 this altitude is very high, but great humidity 

 permits standing pools of water outside of the 

 rivers themselves. Full tables of the distribu- 

 tion of the fishes of Colombia and Ecuador 

 vrill be given in the final complete reports. 

 Arthur Henn 



SPECIAL ABTICLES 



possible factors in the variations of the 



earth's magnetic field 



Newall^ has described a very interesting 

 experiment in which " a lamp flame held 

 under an iron or steel wire (which is in cir- 

 cuit with a galvanometer), so that a short 

 portion of the wire becomes red hot, is made 

 to travel slowly under the wire, and it is 

 found that a current appears in the galvanom- 

 eter, the direction of the current depending 

 on the direction in which the flame travels. 

 Tomlinson^ described a similar experiment at 

 an earlier date than ISTewall. While this cur- 

 rent is described as due to difference in thermo- 

 electric quality between the iron or steel in 

 the magnetic and non-magnetic state, yet it 

 is suggestive of what might happen in the 

 crust of the earth as the sun's rays fall upon 

 its surface and warm it. 



The following simple experiments were 

 carried out with a view to getting more light 

 on the phenomenon of earth currents and 

 their relation to the earth's magnetic field. A 

 board, seventy-five centimeters long and four 

 centimeters wide. Fig. 1, had a shallow rim 

 fastened around it so as to form a tray. At 

 either end was fastened a zinc strip, both of 

 which were in turn soldered to copper wires 

 leading to a galvanometer. In this tray and 

 covering the zinc terminals, a fairly homo- 

 geneous paste of mud was placed about one 

 half centimeter thick. The water used in 

 making up the mud paste was slightly acidu- 

 lated with sulphuric acid to make it a better 



1 Newall, Philosophical Magazine, June, 1888. 

 See also Ewing's "Magnetic Induction," p. 184, 

 3d ed. 



2 Tomlinson, Philosophical Magazine, January, 

 1888, p. 50. 



conductor. When a Bunsen flame was allowed 

 to play on this strip of mud for some little 

 time and then slowly moved in one or the 

 other direction lengthwise of the tray, a cur- 

 rent was set up in the galvanometer, depend- 

 ing upon the direction of the motion of the 

 flame. 



/I _ 



+ 



If the electrode marked B, Fig. 1, was made 

 positive by applying the positive pole of a dry 

 cell to it and the negative pole to A, then the 

 galvanometer gave a deflection to the left. 

 When the flame was slowly moved from A 

 toward B, the deflection of the galvanometer 

 was to the right, and when the motion of the 

 flame was reversed the current was also^i 

 This indicated that the direction of the cur- 

 rent was opposite to that of the burner. 



Suppose now this condition exists in the 

 surface of the rotating earth as the heat rays 

 of the sun falling upon it move from east to 

 west. A current will be set up in the opposite 

 direction, i. e., from west to east, which will 

 locally complete itself on the earth's surface 

 somewhat as shown in Fig. 2. 



It was found that the hotter the Bunsen 

 flame for a given rate of moving, the greater 



