750 



SCIENCE. 



[N. S. Vol. VIII. No. 204. 



students that they put down everything they 

 do, so that it may be judged how it has been 

 done. My views are that when one puts down 

 figures that he did not observe he perpetrates 

 a lie, but that when he puts down figures that 

 do not agree, or carries them too far, he simply 

 makes himself unnecessary labor, but deceives 

 nobody. It is easier to throw off the unneces- 

 sary tigures at the end than to find out 

 what was observed if it is not stated. I pre- 

 sume that this will explain the reason for Mr. 

 Taylor's putting down the readings of the mi- 

 crometers of the cathometer as actually read, 

 that is, to a thousandth of a millimeter. It is 

 not stated that several settings would agree to 

 this amount, but that the figures given are the 

 means of several settings. In the next column, 

 in which the ten-thousandth of a millimeter 

 appears, ' X ' might have noticed, with a little 

 more careful reading, that the last figure is in- 

 variably a five or zero, as each entry is the 

 mean of two. I am not aware what proced- 

 ure ' X ' would adopt in taking means. As the 

 result of all the measurements of the coil, the 

 mean diameter of the coil is presented to seven 

 figures. Now it is not a little singular that in 

 Lord Kayleigh's celebrated paper on Clark 

 cells, in the description of the current- weigher, 

 of which ours is a humble imitation, the mean 

 radius of the two coils is given to exactly seven 

 figures, the results being 'derived from the di- 

 mensions recorded in Professor Maxwell's hand- 

 writing in the laboratory note-book.' Lord 

 Eayleigh did not even think it necessai-j' to un- 

 wind the coil, and it evidently did not occur 

 to him how mortified Maxwell would have felt 

 had he been charged by 'X' with 'implying 

 that his measux'es were made to one part in 

 two millions.' Of course, Mr. Taylor ought to 

 have known that what Maxwell could do he 

 could not. I charge myself with remissness in 

 not having impressed this upon him. 



With regard to the inquiry as to the maker 

 of the cathetometer, ' of a type so extraordi- 

 nary as to justify these figures,' it was a fine 

 instrument by the Societe Genevoise, fastened 

 to the wall, a photograph of which appears in 

 the May-June number of the Physical Review. 

 The levels upon it were by the same makers, 

 whose names there is no motive for concealing. 



With regard to the statement that a degree of 

 change of temperature would probably change 

 the length of the bar by fifteen or twenty 

 thousandths of a millimeter, I find that, ac- 

 cording to Benoit's results for steel, the part 

 of the bar used would change by about two 

 thousandths, so that ' highly perfected methods 

 of determining the temperature' were not 

 used. We have several thermometers capable 

 of reading half-degrees. 



The culmination of'X's' sarcasm is, how- 

 ever, reached in the comment on the computa- 

 tion of the constant, in which it is stated that 

 the last figure stands for one part in thirty mil- 

 lions (italics ' X's '). It is a fact that Lord Ray- 

 leigh only uses seven place figures, Mr. Taylor 

 eight. These he took from Legeudre's tables, 

 throwing off the superfluous figures at the end, 

 where the constant is given to four figures, not 

 eight. If ' X's ' representation in this case be 

 ingenuous I prefer Mr. Taylor's deception. 

 The fact that the computation ' made by other 

 people and a different method ' gives a result 

 differing by one part in a thousand (not in five 

 or six hundred, as stated by 'X') does not 

 throw any discredit on the measurements, but 

 shows that the approximation of the first mode 

 of computation was not sufficient, as is plainly 

 stated. 



I will not stop to dwell on the comments on 

 the weighings. They are of the same sort, and 

 may be answered in the same way. To the last 

 figure, although observed, no weight is to be 

 attached. The supposed ' marvelous skill ' re- 

 quired thus disappears. But to conclude : 

 What, it may be asked, is the use of printing a 

 paper in which the main result, as is frankly 

 stated, is disappointing? To this I may make 

 a brief answer, though it is not touched upon in 

 the communication. The subject of the de- 

 termination of current in absolute measure is 

 one that is now very much in the air. It be- 

 comes important to know what is the best form 

 of measuring instrument, and what is the best 

 method of computation of the constant. To 

 this matter some of the last work of Helmholtz 

 was devoted. The question of computation 

 has been attacked with great vigor by Principal 

 J. Viriamu Jones. An elaborate current-weigher 

 of the Rayleigh type was constructed by the 



