April 28, 1899.] 



SCIENCE. 



625 



produced. Here the effect was caused, not by 

 the low potential of the battery current, but by 

 the much higher potential of the extra current 

 produced by the cons' ant breaks at the lower 

 end of the thread. This was demonstrated by 

 placing a non-inductive resistance of about 1,- 

 000 ohms in parallel with the coherer. This 

 entirely prevented the action. Attaching the 

 ends of the secondary of an induction coil to 

 the coherer gave similar results, as did also the 

 Holtz machine. 



But these threads remained cohering after 

 there had ceased to be a difference of potential 

 between the electrodes. In a vacuum the 

 threads still remained hangitfg, showing that the 

 friction and pressure of the air did not maintain 

 the coherence. Under the microscope the 

 points of contact appear to be fused, and other 

 observers have noticed bright points after cohe- 

 rence is destroyed. That fusion occurs is also 

 shown by the fact that metals having a high 

 melting point give threads of much less tenacity 

 than those with low melting points. Thus 

 platinum and iron give very fragile threads, 

 while tin, lead and aluminum give threads 

 capable of enduring considerable flexure. 



With the Holtz machine threads could be 

 produced only when the machine was run 

 slowly. If it was run too fast the particles 

 would fly back and forth between the elec- 

 trodes of the coherer. The point of the 

 coherer, becoming charged, induces a charge 

 in the particle nearest to it. This causes 

 an electrostatic attraction, the particle flies 

 to the point and, receiving a like charge, is 

 at once repelled. But the instant it comes in 

 contact the points of contact fuse, and if the 

 ■charge on the electrode be small this fusion 

 will be suflicient to resist the tendency to fly off 

 and the filing will remain, becoming a part of 

 the electrode and repeating the action on the 

 next particle. Thus consecutive filings are 

 united and a continuous thread of filings fused 

 together, which connects the electrodes, reduc- 

 ing the resistance greatly. In a Hertzian field 

 the action is precisely similar, the difference of 

 potential here being produced by the action of 

 the Hertz waves. Short threads of filings were 

 obtained, as in the preceding experiments. 



These experiments show that the great re- 



duction in the resistance of the tube is due to 

 the formation of continuous threads of metal 

 connecting the electrodes. They also indicate 

 some of the points which a good coherer must 

 possess. The filings used should be composed 

 of metal not easily oxidized, of small specific 

 gravity and low melting point. The electrodes 

 should consist of points or roughened surfaces 

 of similar metal. Further, it is difficult to de- 

 cohere a thread while a current is flowing, 

 since the induced current at any break tends to 

 bring back the parts to coherence. Marconi 

 avoids this action by the use of a high resistance 

 in parallel with the coherer. The necessity for 

 this resistance can be avoided by having the 

 current through the coherer broken before the 

 tapping occurs. Experiments are being con- 

 tinued in the direction of a practical application 

 of these principles. 



M. H. LocKWOOD, 

 E. B. Wheeler. 

 March 30, 1899. 



two-headed snakes. 



To THE Editor of Science : I am en- 

 gaged in the study and description of two- 

 headed snakes by means of skiagraphy. Al- 

 though I have in hand eight specimens from 

 various museums, I have been unable to locate 

 the Tropidonotus from the Massachusetts State 

 Collection, described by Wyman in the Proc. 

 Bosf. Soc. Nat. Hist., Vol. IX., p. 193, and the 

 three snakes described by Mitchell in the^mer. 

 Jour, of Science, Vol. X., p. 48. 



I write in the hope that one of your readers 

 may be able to help me in my quest of these 

 four specimens, and that I may be informed of 

 any other snakes with this abnormality in 

 American collections, in order that I may make 

 note of, or describe, them in my forthcoming 

 paper. 



KoswELL H. Johnson. 



1727 Cambridge Street, Cambridge, Mass., 

 April 14, 1899. 



DUPLICATION OF GEOLOGIC FORMATION NAMES. 



It was not my intention, in my letter in Sci- 

 ence of March 31st, to discuss the question as 

 to whether certain names of geologic formations 

 conflicted, or to discuss the undesirability of 



