February 23, 1922J 



NATURE 



23^ 



its being granted, but I have always understood 



it it was reserved for the exceptional case of a 



ious scientific investigator whose academic career 



been interrupted after the intermediate stage and 



|ho wishes to resume it at an age when the con- 



itrated study necessary for the passing of the final 

 I.Sc. examination cannot reasonably be required of 

 If there is any reason to fear possible abuse 

 this means of graduating, it would be a simple 

 matter to fix a minimum age-limit for it — say thirty- 

 five — rather than to abolish it. 



The immediate importance of the B.Sc. by research, 

 however, is that it gives the sub-committee an oppor- 

 tunity to exaggerate the number of standards of 

 lesearch which examiners have to keep in their 

 uinds. The report says: "We do not consider it 

 satisfactory that there should be as many as three, 

 ind still less four, degree standards of research." 

 Ihere is no fourth standard. I cannot conceive any 

 examiner recommending the B.Sc. degree for a thesis 

 which would be rejected for M.Sc. if offered by a 

 graduate. On the contrary, I can easily imagine 

 the University making it a rule to accept no thesis 

 for B.Sc. which it would not accept for Ph.D., or 

 even D.Sc, if this means of graduation is reserved 

 for very exceptional cases. The conditions under 

 which alone B.Sc. by research should be granted for- 

 bid the standard from making a fourth with the 

 three standards of post-graduate research degrees. As 

 to those three standards, as an examiner I have not 

 found any difficulty in framing three standards in my 

 own mind or in agreeing upon them with my col- 

 leagues. On the contrary, I find that the introduc- 

 tion of the Ph.D. degree has made it easier to define 

 the standards of the two others. It the M.Sc. by re- 

 search be abolished, the Ph.D. standard will inevit- 

 ably tend to sink, until in a few years it will be 

 equivalent to the present M.Sc. 



I am certainly not speaking for myself alone when 

 I express myself as strongly in favour of the retention 

 of the M.Sc. by research ; but if the University 

 should decide to abolish it, I should very much prefer 

 the abolition to be complete rather than that the 

 degree should be granted by examination. The work 

 of a candidate for M.Sc. by research must neces- 

 sarily consist very largely (in some cases entirely) in 

 a survey of the knowledge already acquired on the 

 subject which he proposes to investigate. This 

 involves the intensive study of original works of re- 

 search possibly going far back into the early history 

 of science and extending into various branches, all 

 connected together bv their bearing on some one 

 problem. In geology, for instance, such an investiga- 

 tion may often include portions of the several 

 branches — petrology, stratigraphy, palaeontology, and 

 economic geology. 



A candidate studying for M.Sc. by examination will 

 be engaged on very similar work, but, its boundaries 

 will be arbitrarilv determined for him by the defini- 

 tion of some particular subject which he chooses from 

 a published list. He will be warned off side-issues 

 that may attract him by the fact that they will not 

 form part of the subject-matter of his examination. 

 He will be much more inclined to rely on text-books 

 than on original papers, and any tendency to run 

 down obscure questions for himself in the literature of 

 the subject or by personal observation and research 

 will be positivolv disadvnntngoous to him, since he will 

 be discovering facts probably unknown to his examiners. 



Examples could easily be found of able investigators 

 whose life-work originated as a side-issue from an 

 early line of study. At the beginning of post-graduate 

 life a man cannot be expected to choose irrevocably 

 his main line of work. 



NO. 2730, VOL. 109] 



As an examiner I am convinced that the Honours 

 B.Sc. stage is the highest at which examinations are 

 of value, except as an altogether subordinate part of 

 the qualifying test. After this stage every incentive 

 should be given to the student to work on lines deter- 

 mined by his particular interests and opportunities, 

 and not by what must be, even when every effort is 

 made to avoid it, an arbitrary pigeon-hole sub-division 

 of the sciences. A, Morley Davies. 



Imperial College, S.W.7, February 9. 



The Accuracy of Tide-predicting Machines. 



I SHOULD like to make a few comments on Mr. 

 Marmer's letter in Nature of February 2, p. 136, as 1 

 was responsible for the tests made on the British 

 machines referred to in your review of " British 

 Research Work on Tides." 



In his last paragraph Mr. Marmer states the various 

 uses that can be made of tide-predicting machines in 

 addition to their normal use. Most of these are quite 

 likely to be well within the capacity of any machine, 

 since relatively small quantities only are involved and 

 the full scale of the machine can be used. But their 

 use in " the elimination from the observed tide of the 

 tide due to a number of constituents " is precisely that 

 which was shown to be undesirable so far as the 

 British machines were concerned. It has been found 

 very advantageous in research work to subtract known 

 constituents from the tidal record and to examine the 

 residue, but for such work it is of prime importance 

 to know that what we have actually removed is exactly 

 what it professes to be. It is not desirable to spend 

 time and energy on the examination of fictitious 

 residues due to machine errors, and it was found that 

 the British machines were subject to systematic errors 

 of about 05 ft. in hourly heights (though not in 

 heights of high and low water), with a spring range 

 of 18 ft. Such errors entirely prohibited the use of 

 these machines. 



It is quite probable that the performance of the 

 British machines can be improved, but the labour of 

 reading the curves will be great. In this respect the 

 U.S.A. machine has a notable advantage, and I 

 should be very glad to know that one could obtain 

 from it hourly heights with an accuracy suitable for 

 research work, say to within 0-05 ft. for a spring 

 range of 30 ft. But in fairness to the British 

 machines, and not with a desire to impeach the work- 

 ing of the U.S.A. machine, I must say that I am not 

 convinced bv the tests recorded by Mr. Marmer. At 

 Hong Kong the spring range of tide is only 45 ft., 

 and if the full powers of the machine have been used, 

 as is reasonable to suppose, then we should expect a 

 j)ro rata error of 04 ft. with a spring range of 30 ft. 

 It is fervently to be hoped that such is not the case, 

 though I must confess that certain comparisons 1 

 have made between direct calculations and U.S.A. 

 predictions show discrepancies of this magnitude, even 

 in high- and low-water heights. Further, the differ- 

 ence in predictions between the U.S.A. machine and 

 one of the British machines is much greater than is 

 to be expected, if it be due to the errors only of the 

 latter. 



It is very noteworthy that the performance of the 

 U.S.A. machine in T022 agrees very well with its per- 

 formance in iqio, indicating that its errors are trulv 

 svstematic: but this is no consolation to a research 

 worker unless he knows what the errors are. It is 

 easy to see that the errors have not anv obvious rela- 

 tionship to the actual tide predicted. The tests illus- 

 trate the difficultv one would have in dealing with the 

 residues, for of the thirty constituents used about half 

 are individuallv less than the error of the machine. 



