February 13, 1902] 



NA TURE 



545 



some materials for a historical introduction, but practi- 

 cally no actual text. The councilor the Sanitary Institute 

 did me the honour of asking me to undertake the prepar- 

 ation of the report. For obvious reasons, the part that I 

 could take in such an enterprise could only be a small 

 one ; it was limited to supervising the work done by two 

 of Mr. Rogers Field's assistants and writing an occa- 

 sional general note. In doing so I have regarded the 

 report as addressed to those interested in the scientific 

 study of pneumatics, and have not hesitated to call 

 attention to the points in which the action of the com- 

 mittee seemed to me to have missed the true scientific 

 bearings of its work. I propose to be equally frank in 

 what I have now to say, but I would not thereby be 

 understood as decrying the accuracy or the value of the 

 experiments. The study of pneumatics has remained 

 undeveloped probably more on account of the lack of 

 accurate experiments and accurate measuring instru- 

 ments than for any other reason. The records of the 

 cowl committee bear ample testimony to the single- 

 ness of purpose of the committee in its desire to take 

 all precautions and bring all conclusions to the test of 

 .accurate experiment, and the records of the experi- 

 ments, with the limits of accuracy clearly apparent, 

 rnust always be regarded as valuable data by which to 

 test theoretical conclusions and an honourable memorial 

 of a worthy effort to enlarge our knowledge of a most 

 intricate subject. 



The object of the committee was to compare, by direct 

 experiment, cowls of different type as agents for producing 

 a flow of air when exposed to wind and for preventing 

 downdraught. The procedure adopted was to have three 

 long vertical tubes projecting from a weather-boarded 

 hut, erected, by permission of the Kew Observatory com- 

 mittee, in the Old Deer Park at Richmond. Each tube 

 had an air meter inserted in it to measure the flow of air ; 

 the lower end was protected from draughts by a box with 

 a silk gauze bottom ; an anemometer gave the velocity of 

 the wind, the direction of which was also recorded. The 

 cowl was mounted on the middle of the three pipes, and 

 the flow through that pipe was referred to the mean of 

 the flow through the two external pipes. The tests of 

 downdraught were of various descriptions that need not 

 be referred to here. 



With the comparison of cowls of recognised shapes there 

 was associated an endeavour to ascertain the effect of 

 different modifications of the orifice of a pipe, and for 

 this purpose a number of such modifications were con- 

 structed and examined. Such modifications, some of 

 which were very elaborate, are called terminals. 



The report is divided into six parts : — 



Part i. consists of extracts from introductory papers 

 prepared by various members of the committee or from 

 the correspondence preserved by the committee. 



Part ii. contains an account of the testing of the air 

 meters, and will be found to give a considerable amount of 

 information of very great practical utility concerning the 

 action of those instruments. 



Part iii. gives the results of Mr. Rogers Field's investi- 

 gation of the action of the anemometers employed. The 

 instruments principally considered are a Robinson ane- 

 mometer of the standard Kew pattern and a miniature 

 instrument of the same type with one-inch cups. The 

 chapter does not add much to the solution of the general 

 problem of determining true wind velocity from the 

 reading of a Robinson instrument, but it does show what 

 kind of difficulties a careful and conscientious experi- 

 menter is likely to meet with if he sets about using such 

 an instrument for the determination of true wind velocity. 

 It also shows by some very useful diagrams the relation 

 between the wind velocity and the flow up a vertical open 

 tube over the end of which the wind passes. 



Part iv. gives an examination of the degree of accuracy 

 with which the mean of the flow in two outside pipes can 



NO. 1685, VOL. 65] 



be regarded as equivalent to the flow up the middle pipe. 

 In some ways the results of this section are the least in- 

 teresting of the whole number, because the observed 

 differences which the committee attemped to resolve 

 must be regarded as due to local circumstances which 

 would probably not be reproduced in a repetition of the 

 experiments under somewhat different conditions ; but 

 they are an essential part of the work of the committee ; 

 they show the limits of accuracy of the measurements 

 under the prescribed conditions. Probably an experi- 

 menter with long experience of a laboratory might have 

 been content to recognise after a few experiments a 

 certain margin of experimental error as incidental to the 

 method, and have left the cowl results with that margin 

 of error, or have selected conditions which gave the 

 least experimental differences ; but the committee seemed 

 unwilling to write off an experiment as subject to a 

 certain margin of error until it had fully probed all the 

 causes of error. 



Part V. gives the results for terminals, as defined above. 

 The results which can best be generalised are those which 

 are represented in the report by what are known as " hill 

 curves," by which are to be understood curves represent- 

 ing results obtained from a series of consecutive experi- 

 ments upon terminals varying by the gradual extension 

 of some particular dimension. 



Part vi. gives the results for cowls. 



A consideration of the whole report gives rise at once 

 to curious reflections. If an apostle of higher education 

 were looking for an example of the importance of recondite 

 theoretical study to matters of practice, of the necessity 

 to practical life for the academic professor in any subject, 

 not as an exponent of the facts of the subject, but as a 

 student and investigator of its abstract laws, he could not 

 wish for a better example than that furnished by this 

 report. « 



The original committee was appointed to settle an 

 apparently simple practical matter, namely, which was 

 the best among a number of cowls exhibited at Leaming- 

 ton in 1876 in competition for a prize or certificate. The 

 matter was apparently confined to the region of practice, 

 and it was at first assumed that only a few experiments 

 were needed to settle the points in question. Experi- 

 ments were made, and the committee reported to the 

 effect that no cowl at all was as good as any ; and the 

 award went in consequence to no cowl at all. But this 

 did not by any means satisfy all concerned, and numerous 

 complaints were made as to the experiments and the way 

 of conducting them. 



Thereupon the Sanitary Institute appointed Sir Douglas 

 Galton, Mr. Rogers Field and Mr. VV. Eassie — the last- 

 mentioned was subsequently succeeded by Mr. J. Wallace 

 Peggs — to be a committee to conduct further experiments. 

 They set out to repeat the experiments with such pre- 

 cautions that their results should be accepted as final. 

 They worked with unremitting labour and at no small 

 expense ; tested to the uttermost every instrument em- 

 ployed in the investigation, and analysed all the condi- 

 tions that might affect the results. They acted through- 

 out upon the apparently simple practical principle that 

 they could find out which was the best cowl if they could 

 find out, for a certain strength of wind, which carried 

 most air up a three-inch or a six-inch pipe, and up to 

 what angle of tilt it could be set in various circum- 

 stances without suffering downdraught. 



It certainly cannot be regarded as a fault in the com- 

 mittee that the subject had not been effectively worked 

 at by some academic professor or student of experimental 

 philosophy curious to learn, not which cowl should have a 

 prize, but the general laws of flow of air through any cowl. 

 They were capable, practical men, and naturally attempted 

 a direct experimental answer. It may be true that in nine 

 cases out of ten the best way of getting a practical answer 

 to a scientific question is to set practical men to find it, 



