SCIENCE. 



FRIDAY, OCTOBER 19, 1883. 



PliECISrON OF OBSERVATION AS A 

 IS RANCH OF INSTnUCTlOxY. 



We sometimes find the philosopher ciivving 

 the physicist because the results sought b}- the 

 latter can be expressed exactly. In so doing, 

 he doubtless overlooks the fact that all quanti- 

 tative work has to be regarded as giving mere- 

 ly an approximation to the truth. Numerous 

 and refined as are the precautious adopted, 

 the careful experimenter must admit that his 

 rneasuremeuts contain errors whose sources are 

 more or less hidden. Success will lie, not in 

 ignoring this, but in recognizing it, and studi- 

 ously- avoiding anj- unwarranted claim to accu- 

 raej". His investigation maj' establish some 

 law beyond a reasonable doubt. This law may 

 be expressible in exact terms ; but, so far as 

 the direct quantitative results are concerned, 

 he must give up, once for all, the popular 

 notion of exactness. He must admit that the 

 work merely shows it to be reasonable to as- 

 sume the truth to lie somewhere between two 

 limits, respectivelj- greater and less than the 

 one magnitude which he names. Desiring 

 that these limits shall fall as near together as 

 practicable, he will stud^- the obsen-ations for 

 internal evidence of the precision attained ; 

 but any appearance of accuracy' greater than 

 might rcasouabl3- be expected will often cause 

 him more uneasiness than would a greater ap- 

 parent eiTor. Until the extent of the error is 

 recognized, and found to be in harmony' with 

 experience derived from other similar observa- 

 tions, a cautious ol)ser\er will not be confident 

 of the result reached. 



As the scientific professions are currently 

 taught, it is possible to get a fair training 

 (as a chemist, engineer, or electrician, for 

 example) without properly appreciating the 

 practical limitations to precision. He maj- 

 instinctivel}' acquire correct notions of the 



No. 37.-1883. 



performances of the instruments he most ifre- 

 qucnlly uses ; but let a new operation, with a 

 ditferent instrument, be required, and he will 

 too often develop the wildest notions as to the 

 great accuracy attainable by the use of suflB- 

 cicnt care. What is needed in our professional 

 courses is systematic instruction in the general 

 science of planning, conducting, and discuss- 

 ing observations, accompanied with adequate 

 practice. This should be given as early as 

 the student is fitted to profit by it, in order 

 that the subsequent practical training in special 

 branches ma}' have a firm foundation. 



The field for such instruction is ample. A 

 good routine observer is one who, being in- 

 formed of the accurac}' desired at each step, 

 is able to take just care enough to attain it, 

 without wasting time and energy in uselessly 

 perfecting certain parts of the work. Our 

 professional observer must add to this the 

 good judgment which is able to discover the 

 relative accuracy required in different parts of 

 a complex observation, and to decide how ac- 

 curate to aim to make a single performance of 

 the whole. In general, he will seek to avoid 

 errors which usually occur in a single direc- 

 tion ; but he will not alwa3-s take the greatest 

 care to avoid errors which are as liable to be 

 negative as positive. Life is short. Time, 

 to most, is money ; and the ability to repeat 

 an observation will often depend upon the 

 ability to do it quickly. Moreover, in many 

 cases, mere lapse of time allows additional 

 errors to enter. Having avoided the larger 

 errors, he will therefore seek to eliminate 

 the effect of the smaller by repeatedly per- 

 forming the work. Recognizing, then, the 

 importance of reasonable speed, he will allow 

 rough measures of certain quantities, pro- 

 vided the final error of the complex operation 

 is not there bj- appreciably affected. All this 

 calls for a clear understanding of the causes 

 of error, and an ability to reason out their 

 effect upon the result. The knowledge of the 



