628 



SCIENCE 



[N. S. Vol. XXXI. Jv'o. 799 



of terms employed. The plates appear to be 

 carefully drawn and are artistically excellent. 

 A careful examination of the book leaves a 

 delightful impression upon the mind. It is 

 in many respects a model of monographic 

 treatment, and the editor, Dr. A. E. Shipley, 

 is quite right in saying that whereas 



Dr. David Sharp in the Fifth Volume of the 

 " Cambridge Natural History " states : " The 

 classifieation of the earwigs is still in a rudi- 

 mentary state." . . . Burr's work will cause the 

 deletion of this sentence if a new edition of Dr. 

 Sharp's volume be called for. 



The author of the work intimates that he is 

 engaged in preparing upon the same lines an 

 account of the Dermaptera of the entire world. 

 The appearance of such a work will certainly 

 be welcomed, and the present reviewer hopes 

 that the learned author may be spared in 

 health to complete it at no distant day. 



W. J. Holland 

 Caenegie Museum, 

 March 15, 1910 



/ 



SPECIAL ARTICLES 



CANAL-RAY EFFECTS IN OPEN AIR DISCHARGE 



In a paper recently published' the writer 

 has shovTn that the positive luminescence in 

 a Geissler tube is due to a progressive ioniza- 

 tion of the air column, and that this ioniza- 

 tion begins at the anode wire. In a long tube 

 like that used by J. J. Thomson, this ioniza- 

 tion may extend over a distance of fifteen 

 meters. 



Since the publication of the paper, evidence 

 has been secured on photographic plates, 

 showing that a disruptive spark discharge in 

 open air can not be produced, until such 

 ionization, originating at the anode terminal, 

 has reached the negative terminal. 



Confirmation of this conclusion may be ob- 

 tained in the manner now to be described. 

 We have used a large eight-plate influence 

 machine. 



Small spark-knobs are so adjusted that a 

 torrent of loud sparks passes between them. 

 Hang midway between the knobs a sheet of 



'■ Trans. Acad, of 8c. of St. Louis, Vol. XIX., 

 No. 1. 



copper. It is suspended on long silk threads, 

 its plane being at right angles to the line 

 joining the knobs. The sparks can not now 

 be made to pass. A column of positive 

 luminescence joins a positive terminal and a 

 copper plate, but the cathode half of the gap 

 is dark. A glass rod interposed in the positive 

 luminescence casts a shadow on the side 

 turned away from the anode. The shadow is 

 not bounded by right lines, as is the case in 

 rarefied air, where the mean free path is 

 great. When the rod is held near the copper 

 plate, a shadow is, however, cast on the plate. 

 If the plate is moved to a parallel position 

 near the negative terminal, a torrent of 

 sparks passes through the plate. If moved in 

 the opposite direction, until it makes contact 

 with the positive knob, no sparks will pass in 

 any position of the plate. A negative inflow 

 to the edges and corners of the plate is now 

 taking place, as is shown by brush "dis- 

 charges," but the ionization effects are dis- 

 persed in such a way that the conducting 

 channel or channels through the air do not 

 lead to the negative terminal, and no spark 

 can pass unless the spark gap is made shorter. 



We have here a clear explanation of the 

 reason why the spark length is greater, when 

 the positive terminal is a small knob than 

 when it is a large one. 



A small windmill was placed in the positive 

 luminescence, with its plane of rotation at 

 right angles to the discharge. The vanes were 

 of thin mica sheet. The diameter from tip to 

 tip of the vanes was 8 cm. The vanes were 

 mounted on a hub of hard rubber having a 

 shaft of vulcanized fiber, and turning on 

 pivots of fiber or glass, mounted in hard rub- 

 ber. The vanes rotated in a direction which 

 showed that the air was drifting away from 

 the positive terminal. As nearly as could be 

 estimated, the rotation was such as was pro- 

 duced by carrying the mill through still air 

 with a velocity of 1.5 meters per second. 



All of the results described are produced 

 when the negative terminal is grounded. 



These phenomena show that in all proba- 

 bility an X-ray tube will be much less likely 

 to suffer puncture, if its cathode is grounded. 



