SCIENCE. 



171 



THE CO-EFFICIENT OF SAFETY IN NAVIGA- 

 TION* 



By Prof. W. A. Rogers, of Harvard University. 



It is customary among engineers and architects, in 

 making allowance for the strain to be borne by any part 

 of a structure, to assign to the materials used, a strength 

 sufficient to withstand a strain somewhat greater than 

 the structure is ever likely to be subjected to. By experi- 

 ment it can be found, for example, what is the "breaking 

 load" of a wooden or iron beam of given dimensions, 

 and an empirical law is establ shed which will give us 

 approximately the breaking load of any beam, when we 

 know the dimensions, material, etc.; but in order to 

 cover all possible differences which may exist in various 

 beans, a Coefficient of Safety is either introduced into 

 the formula itself, or is applied to the result obtaiued 

 from the formula. This co-efficient should be large 

 enough to cover, not only the largest possible deviation 

 between experiment and theory, but also to meet all un- 

 foreseen emergencies, such as time and age inevitably 

 bring. 



Passing now to the consideration of the term " Co- 

 efficient of Safety," as applied to navigation, it is our 

 object to find the limits within which, under ordinary 

 circumstances, a vessel can be located at sea, and then 

 adduce some considerations which will enable us to form 

 an intelligent judgment in regard to the range of error to 

 which observations are liable. The quantity wanted is 

 the average number of miles error in latitude and longi- 

 tude, which we may fairly charge upon a single observa- 

 tion at sea, under ordinary circumstances. We have 

 then to find the co-efficient by which this number must 

 be multiplied in order to secure absolute safety, as far as 

 safety depends upon human means and exertions. 



By an examination of the " British Wreck Register" 

 and the official inquiries made into the causes of disasters 

 at sea, it will be seen that the ratio of loss compared 

 with the increase of tonnage afloat, has for many years 

 been steadily increasing. This inquiry is, therefore not 

 an idle one. It is our purpose to examine only those 

 causes of wrecks which, in a measure, seem to have es- 

 caped attention in official investigjtions. They are : — 



I. — Wrecks produced by causes clearly beyond human 

 control. 



II. — Wrecks resulting directly or indirectly from over- 

 insurance. 



III. — Wrecks caused by the deviation of the compass. 



IV. — Wrecks caused by errors of observation at sea. 

 The first inquiry is an important one, since, if we can 



find how many wrecks are beyond human control, we 

 ascertain, at the same time, how many are within human 

 control. The method of investigation is by the examina- 

 tion of records of Courts of Inquiry for twenty years. Be- 

 tween 1785 and 1 81 3 no less than eight British ships were 

 either wholly or partially disabled by lightning. Of course, 

 vessels lost and never heard from should be added to this 

 list. Between 1864 and 1869 we find from the insurance 

 records that 9999 sailing vessels and 589 steamers, or a 

 total of 10,588, were wrecked. Of this number, the end 

 of 846 is entirely unknown, or one-eighteenth of the 

 whole number. It is probable, therefore, that seven out 

 of ten wrecks occur from preventable causes. 



In regard to the second head, it is certain that more 

 insured than uninsured vessels are lost, and in not a few 

 cases it has been possible to convict shipowners of pur- 

 posely destroying their vessels. 



The compass problem is an intricate one, and has never 

 been fully solved, though the researches of Flinders, 

 Barlow, Scoresby, Airy and Harkness have done much to 

 convert great uncertainty into tolerable certainty. The 

 first observations on the variation were by Bond, in 1668. 



* Abstract of a piper read before the Nav.il Institute of Annapolis. 

 Prepared under author's direction. 



It is well known that the variation of the needle is very 

 irregular. There are yearly, monthly and diurnal inequal- 

 ities, the diurnal variation being discovered by Graham, in 

 1722. But the complexity of the problem does not stop 

 here. The tendency of the present time is to build iron 

 ships, and all ships now have more or less iron in their 

 construction. These ships become, to a greater or less de- 

 gree, themselves great magnets. In wholly iron ships the 

 uncorrected deviation of the needle often amounts to 50°, 

 thus rendering it utterly useless. The Admiralty Law in 

 regard to "swinging" for the variation of the compass is 

 a very clear statement of the case. It reads as follows : 



" As the deviation or error of the compass caused by 

 local attraction of the ships becomes changed in amount 

 by any change in the ship's geographical position, and 

 may be entirely reversed in its direction by the ship's pro- 

 ceeding from the Northern to the Southern Hemisphere 

 it is to be invariably tested by azimuth and amplitude 

 observations at sea, and the ship is to be swung for ascer- 

 taining the change of error on arrival at a foreign station, 

 and also once a year, and the same is to be inserted in 

 the logbook and sent to the Admiralty with the quarter^ 

 return for December." 



The next important discovery in this connection was 

 by Barlow, who found that all the influence of iron bodies 

 exerted on the compass resides on the surface. This dis- 

 covery paved the way for Airy's method of correcting 

 compasses, which is by swinging the ship in the usual 

 way, and then correcting the local attraction of the ship 

 by means of permanent magnets of soft iron conveniently 

 placed with respect to the compass. But the most im- 

 portant discovery was made by Dr. Scoresby. He found 

 that every iron ship is itself a magnet, and that it gets 

 its magnetism while building by the inductive magnet- 

 ism of the earth, the poles of the ship's magnetism de- 

 pending on the position of the building yard and the 

 direction of the keel in construction. Dr. Scoresby made 

 the voyage of the world in the Royal Charter, to test his 

 theory, and found it fully confirmed. Before starting, his 

 compasses were corrected by Airy's plan. On arriving 

 at Melbourne, it was found that a complete inversion 

 of the ship's magnetic polatity had taken place. Every 

 stanchion, every standard, every davit, every mass of iron 

 about the deck had in its upper surface acquired a North- 

 ern, instead of a Southern polarity, and the starboard 

 compass had lost nearly one-half of its original errors. 

 On returning to the place of starting in the Northern 

 Hemisphere, and swinging the ship, it was found that a re- 

 inversion had taken place, but the compasses did not 

 quite return to their original deviations, but retained a 

 fraction of their errors. It has since been found that 

 these changes are much greater in steam than in sailing 

 vessels, as shown by observations on board the Vulcan 

 (steam), and the Pandora (sailing). 



In 1852-53, Dr. Scoresby, in a paper before the British 

 Ass^iation, showed that there is a sensible difference in 

 the deviation before and after steam is up. It has been 

 said that the compasses of a steam vessel, when light and 

 running before the wind, with a high sea, are practically 

 useless. 



More recent experience has shown that the magnetism 

 of an iron ship does not attain its normal condition till 

 some twelve months after launching, and that for some 

 time the variation is very irregular. In the Great Eastern 

 a fixed compass changed its deviation nearly 3 points in 

 the first 9 momhs of service. The observations of Prof. 

 Harkness on board a monitor seem to be conclusive on 

 this point. 



We come now to the consideration of the fourth point. 

 As early as 1 598, Spain offered a reward of 100,000 

 crowns for the discovery of a correct method of finding 

 the Longitude at sea. The States of Holland, at an 

 early date, offered a reward of 100,000 florins, and France 

 a reward of 100,000 livres. In 1714 the British Govern- 

 ment offered a reward of ^10,000 to any one who should 



