OcTOflKR (i, 19-2 1 



SCIENCE 



375 



mental position of the beginner. An unex- 

 pected result is obtained, an analysis fails to 

 agree, and the supervisor, out of his long ex- 

 perience, can explain the anomalj' at once, and 

 generally does so. If the pupil is to derive 

 any real benefit from his difficulties, his ad- 

 viser must for the moment place himself in 

 the position of one equally puzzled, and must 

 lead his collaborator to sum up the evidence 

 and arrive at the correct conclusion for him- 

 self. The policy thus outlined is, I believe, 

 sound, but it makes severe demands on patience, 

 sympathy, and, above all, time. 



Research supervision, if conscientiously 

 given, involves the complete absorption of the 

 director's enei-gy and leisure. There is a rich 

 reward in seeing pupils develop as independent 

 thinkers and workers, but the supervisor has to 

 pay the price of seeing his OTvn research output 

 fade away. He will have more conjoint papers, 

 but fewer individual publications, and limita- 

 tions will be placed on the nature of his work 

 by the restricted technique of his pupils. 



I have defined a high standard, almost an 

 ideal, but there is, of course ; the easy alterna- 

 tive to use the technical skill of the graduate 

 to carry out the more laborious and mechanical 

 parts of one's own researches, to regard these 

 young workers as so many extra pairs of 

 hands. I need not elaborate the outcome of 

 such a policy. 



There is another temptation, and that, in an 

 institution of univeraity rank, is for the pro- 

 fessor to leave research training in the hands 

 of his lecturers, selecting as his collaborators 

 only those workers who 'have passed the ap- 

 prenticeship stage. This, I am convinced, is a 

 mistake. liTothing consolidates a researdi 

 school more fli'mly than the feeling that all 

 who labor in its interests are recognized by 

 having assigned to them collaboratoi's of real 

 ability. 



I am not yet done with the professor and his 

 stafi', for they will have oiier matters to attend 

 to if i-eseareh schools are to justify their ex- 

 istence and to do more than add to the bulk of 

 our jommais. In many cases it will be found 

 that the most gifted of the young workers 

 under their care lack what, for want of a better 

 expression, is known as "general culture." 



Remember, these graduates have just emerged 

 from a period of intensive study in which 

 Chemistry and the al'Hed sciences have absorbed 

 most of their attention. For their own sake 

 and in the interests of our subject, they must 

 be protected from the criticism that a scientific 

 education is limited in outlook and leads to a 

 narrow specialism. The research years are 

 plastic years, and many opportunities may be 

 found in the eoui-se of the daily consultations 

 "to impress upon the studerit that there is 

 literature other than the records of scientific 

 papers, and music beyond the range of student 

 songs." I mention only two of the many 

 things which may be added to elevate and refine 

 the research student's life. Others will at once 

 occur to you, but I turn to an entirely different 

 feature of research training, for which I make 

 a special plea: I refer to the inculcation of 

 business-lilie methods. You vail not accuse me, 

 I hope, of departing from the spirit of scholar- 

 ship or of descending into petty detail, but my 

 experience has l)een that research students re- 

 quire firm handling. Emancipated as they are 

 from the restrictions of undergi'aduate study, 

 the idea seems to prevail that these workers 

 ought to be excused the rules which usually 

 govern a teaching laboratory, and may there- 

 fore work in any manner they choose. It re- 

 quires, in fact, the force of a personal example 

 to demonstrate to them that research wo!'k can 

 be carried out with all the neatness and care 

 demanded by quantitative analysis. Again, in 

 the exercise of their new freedom wung col- 

 laborators are inclined to neglect i-ecording 

 their results in a manner which secures a per- 

 manent record and is of use to the senior col- 

 laborator. As a rule, the compilation of results 

 for publication is not done by the experimenter, 

 and a somewhat elaborate system of records 

 has to be devised. It should be possible, 

 twenty years after the work has been done, to 

 quoite the reasons which led to the initiation of 

 each experiment, and to trace the source and 

 history of each specimen anah-zed, or upon 

 whidi standard physical constants have been 

 deteiinined. I need not enter into detail in 

 tliis connection beyond stating that, although a 

 system which secures these objects has for 

 many years been adopted in St. Andrews, con- 



