452 WRITINGS OF JOSEPH HENRY. [1875 



criticism or desire to under- value the results he has obtained, 

 some of which are highly important. Our only object in 

 our remarks and in our investigations is the establishment 

 of truth. 



The determination of the question as to the cause of the 

 abnormal phenomena of sound we have mentioned, and the 

 discovery of new phenomena — are matters not merely of ab- 

 stract scientific interest, but of great practical importance, 

 involving the security of life and property, since they in- 

 clude the knowledge necessary to the proper placing of fog- 

 signals and the instruction of mariners in the manner of 

 using them. 



The hypothesis we have adopted, — that of the change of 

 direction of sound by the unequal action of the wind upon 

 the sound-waves, is founded on well-established mechanical 

 principles, and offers a ready explanation of facts otherwise 

 inexplicable. It is also a fruitful source from which to 

 deduce new consequences to be verified or dis-proved by 

 direct experiment. It would however ill become the spirit 

 of true science to assert that this hypothesis is sufficient to 

 explain all the facts which may be discovered in regard to 

 sound in its application to fog-signals, or to rest satisfied 

 with the idea that no other expression of a general principle 

 is necessary. An investigation however to be fruitful in 

 results must as a general rule be guided by a priori concep- 

 tions. Hap-hazard experiments and observations may lead 

 to the discovery of isolated facts, but rarely to the establish- 

 ment of scientific principles. There is danger however in 

 the use of hypotheses, particularly by those inexperienced 

 in scientific investigations, that the value of certain results 

 may be over-estimated, while to others is assigned less weight 

 than really belongs to them. This tendency must be guarded 

 against. The condition of the experiment must be faith- 

 fully narrated, and a scrupulously truthful account of the 

 resuhs given. While we have used the above-mentioned 

 hypothesis in the following investigations as something 

 more than an antecedent probability, we have not excluded 

 observations which mav militate against it, and we hold 



