for the purposes of Illumination in Lighthouses. 67 



Admitting, however, as must be done, the absolute necessi- 

 ty of improvement in this point, it may be asked, How are 

 strong lights to be procured ? The answer to this is by no means 

 difficult. In using reflectors, we cannot by any union of a num- 

 ber, enable them to penetrate a fog, for twenty Argand burners, 

 placed separately, will disappear nearly at the same distance as 

 one ; but by the introduction of lenses, we can adopt various me- 

 thods of obtaining ten times the light in hazy weather. Some of 

 these methods have been already described ; but another may be 

 mentioned, which is suited only to short distances. In place of 

 having only one large lamp in the focus of the lens, we may suf- 

 round it with Jive or six of the same size. All of them, but one,, 

 will be out of the focus, and they will therefore form slightly di- 

 verging, and slightly converging, columns of light ; but as the 

 distance through which they are required to penetrate is neces- 

 sarily small, they will all add powerfully to the intensity of the 

 main beam, and cause it to penetrate through a considerable 

 tract of hazy atmosphere *. 



The circumstances of the case, however, seem to demand even 

 a more powerful light than can be obtained from oil or gas. Many 

 years ago, Sir WILLIAM HERSCHEL suggested the idea of using 

 in lighthouses the powerful, and almost unsupportable, light de- 

 veloped during the deflagration of charcoal by galvanic action. 

 The suggestion scarcely excited notice, from the enormous expence 

 of maintaining such a light, and from the difficulty of applying 

 it to reflectors ; but though it would be extravagant and unne- 

 cessary to maintain such a light for common occasions, there 

 would be no absurdity in its occasional exhibition, when all other 

 means of illumination fail. 



* If gas were used, we might, on such occasions, employ a burner ten inches in 

 diameter, and having many concentric flames. ', -, 



i2 



