520 REPORT— 1902. 



What. can w e do to mend matters? If we wish to prepare the way for the 

 gradual substitution of a better system, we should have someone responsible for 

 the continuation of the present one. For this purpose it should be recognised that 

 the head of the Meteorological Office is something more than a secretary to a board 

 of directors ; also that he is appointed to conduct meteorological research and not 

 to sign weather forecasts. The endowment of meteorology should mean a good 

 deal more than the endowment of the telegraph office whicli transmits the obser- 

 vations. Terrestrial magnetism and atmospheric electricity are at present looked 

 after by institutions already overworked in other directions, and should be handed 

 over to an enlarged department of meteorology. 



Seismology in this country at present depends on the private enterprise and 

 enthusiasm of a single man, and as long as Professor 3Iilne is willing to continue 

 his work we cannot do better than leave it with him, but some permanent pro- 

 vision will ultimately have to be made. 



An improved organisation such as I have sketched out would do good, but 

 could only very slowly overcome the accumulated inertia of ages. I should prefer 

 a more radical treatment. Organisation is good, but sometimes disorganisation 

 is better. 



Most earnestly do I believe that the subjects of meteorology and terrestrial 

 magnetism, and possibly also of atmospheric electricity, could be most quickly 

 advanced at the present moment if all observations were stopped for five years, and 

 all the energy of every observer and computer was concentrated on the discussion 

 of the results already obtained and the preparation of an improved scheme of obser- 

 vation for the future. When we have made up our minds what to do with the 

 observations ; when we have actually done it ; when we know where our present 

 instruments require refining or supplementing ; and especially when we have found 

 out whether we have not spent much time and trouble on unnecessary detail, then 

 the time will have arrived for us to draw up an economical, sufficient, and efficient 

 scheme of observatioDS. At present we are disinclined to discontinue observa- 

 tions, though recognised as useless, for fear of causing a break. We make ourselves 

 slaves to so-called 'continuity,' which is important, but may be, and I believe is 

 being, too dearly purchased. 



There are no doubt some, though probably not very many, observations which 

 it is necessary to carry on continuously over long periods of time. But at present 

 we are groping in the dark, and go on observing everything, and always in the 

 hope that at some time the observations may prove useful. Our whole point of 

 view in this respect wants altering. AVe should Hx on our problem first and then 

 provide the observations which are necessary for the solution of the problem. If 

 for some definite purpose it is advisable to establish special observatories, let us not 

 be afraid to do so, but especially let us not be afraid to discontinue the observatories, 

 when the purpose for which they have been established has been served. Let us 

 restrict, in the first instance, the secular observations to the smallest number, and 

 concentrate our attention, for short periods of time, on some special question. Let 

 us have, for instance, two or three years of thunderstorm observations, all countries 

 joining in concentrating their energies to the elucidation of all the various 

 features of their phenomena. When that is accomplished it will probably be found 

 that thunderstorms may be left to shift for themselves for a while, and attention 

 directed to some other matter. The whole question of lunar influence on 

 meteorological phenomena might be settled in a comparatively short space of time 

 if the civilised countries of the world could agree to record all observations during 

 a few years according to lunar instead of solar coordinates. Other problems will 

 readily suggest themselves to you, and several might possibly be dealt with 

 simultaueouslj'. 



The great reform I have in view is this : Before you ob.serve make sure that 

 your observations will be useful and will help to answer a definite question. 



Though my frankly outspoken criticisms may not command universal assent, I 

 hope you will agree that there is some foundation for them, and if so, the time 

 is obviously not well chosen when observational science can be separated from its 

 mathematical and experimental sisters. We hope that cosmical physics majr 



I 



