208 



SCIENCE. 



[Vol. XV. No. 373 



by which this number must be multiplied in order to provide 

 for every contingency of danger is 3.5." 



Having called attention to the accuracy, or rather the inac- 

 curacy, with which a ship's position is found at sea, I wish to 

 make a few suggestions which have occurred to me in the 

 course of my professional work, and which ought, perhaps, to 

 be considered by marine authorities. 



In each of our large seacoast cities the Maritime Exchange is 

 the principal place where ship-masters, ship-owners, shippers, 

 marine underwriters, and other men interested in ocean com- 

 merce, meet for the purpose of arranging matters connected 

 with the transportation by sea, and it is through these ex- 

 changes that the plan I have to propose could be most readily 

 carried out. 



"What I have to propose is this: that there should be an 

 ■enforced frequent inspection of the instruments used by ships' 

 ■officers in determining the positions of vessels at sea.' "Whether 

 it would be necessary to have a law enacted by legislation, or 

 whether the marine insurance companies would demand such 

 an inspection before issuing policies, or whether the Maritime 

 Exchanges would take up such a work on their account, I can- 

 not more than surmise. I think likely that the former would 

 be necessary ; but, as the details of the work would be about 

 the same in any case, I have assumed that the members of the 

 •exchanges would be willing to carry on the work without any 

 * 'pressure' ' from outside. 



Let, then, the maritime exchanges of each port adopt a rule 

 i;hat the sextants, compasses, and chronometers of all vessels 

 with which the exchange has dealings shall be inspected every 

 three months, or every six months, or at the close of each 

 voyage, or whatever length of time may be considered 

 sufficient. Let each exchange request the secretary of the navy 

 to assign to duty at its port a sufficient number of officers and 

 men to carry out this inspection. That this last is possible, 

 is shown by the fact that for some years officers of the navy 

 liave been assigned to duty at several of these exchanges for 

 the purpose of collecting observations made on shipboard 

 (the United States Signal Service was also represented there 

 in connection with meteorological matters) . Probably one 

 lieutenant and two subordinate officers would be sufficient to 

 do all the work necessary for the port of New York. As to the 

 details of such a work as is proposed, there is only space to 

 mention enough to show the necessity of it, and to show its 

 firacticability. 



Upon the master of a ship reporting his arrival at the 

 exchange, let the naval officer at once take steps to inform 

 himself of the condition of the sextant, chronometer,'^and com- 

 passes carried by the vessel. 



The sextant is very easily thrown out of adjustment; and, 

 the errors being determined with great difficulty without 

 proper apparatus, they are usually left for the maker to 

 re-adjust. The result is, that in many cases a sextant is used 

 until it is found to be utterly worthless, and then only is it 

 taken to be repaired. 



Let each ship-master be required to show a certificate of 

 ■examination of his sextant; this to hold good for a stated 

 time, and to be furnished free of charge by the exchange. It 

 would not be such a laborious task as it may seem, to examine 

 the sextants; for, if a Neumayer sextant stand (a description 

 is given in the "Archiv" of the Deutsche Seewarte, Jahrgang 

 I. 1878, p. 16) be mounted in a convenient place, it would 

 take the operator but a few minutes to detect the errors in 

 adjustment of a sextant placed thereon, and he could even 

 undertake any minor re-adjustment; but for any serious fault 

 in the instrument it would have to be corrected by an instru- 

 ment-maker. 



The sextants belong to what are known as constant instru- 

 ments, and may keep unchanged for a long time ; but chronom- 

 eters are variable instruments, and have to be compared with a 

 standard as frequently as possible. This difference in the 

 instruments is so easily and generally recognized, that, while 



» This proposal is by no means a Dew one, but some points wliich I shall 

 jnention in connection with it I have not seen mentioned before. 



the sextants are so seldom inspected, the chronometers, on the 

 other hand, have been taken to chronometer-shops at the end 

 of each voyage, no matter bow short, to have the rate of 

 change determined, and to have the error noted just before the 

 departure on the next voyage. I say "have been," because, 

 much to my astonishment, the leading chronometer-maker of 

 Boston informed me some years ago, and shortly after the 

 Boston time-ball had been established, that this time-ball had 

 about ruined the chronometer-rating business, and that most of 

 the ship-masters (especially those of foreign vessels) rated the 

 chronometer on shipboard by observing the fall of the time-ball 

 at noon of each day. As these time-balls have now been estab- 

 lished in the principal ports all over the world, it is safe to 

 conclude that this change in the methods of rating chronom- 

 eters is universal. I consider that this is a step backward in 

 the progress of maritime science, and that there is much less 

 security against navigators getting out of their "reckoning" 

 than existed before this change of method took place. I will 

 state briefly my reasons for thinking this. 



1. A chronometer needs frequent inspection by an expert 

 chronometer-repairer in order to see that all parts of the instru- 

 ment are in good order. By the old method this could be done 

 every voyage when the chronometer was being rated ; but now 

 it is probable, that, in cases where a time-ball can be observed, 

 these instruments will not be submitted to the repairer until 

 after experience has shown it to be in a decidedly bad condition. 



3. A chronometer cannot possibly be rated as accurately by 

 observations of a time-ball as by the means employed in the 

 regular chronometer-shops. In the former case, observations 

 cannot be made on days with rain or fog; it is quite probable 

 that the same person will not be able to make the comparison 

 for the whole succession of days, and a personal difference of 

 half a second or more may occur between two observers ; the 

 error in observing the drop of a time-ball is probably not less 

 than half a second, and many observers would not get the time 

 closer than a second; the time-ball is usually dropped only once 

 daily, so that any error in observing it cannot be detected; 

 the observer on shipboard can have no idea of the errors and 

 the daily changes in the standard clock by means of which the 

 ball is dropped, and must rely implicitly on the accuracy of 

 this signal during spells of cloudy weather, when, as well 

 known, the standard clocks of the best astronomical observa- 

 tories may be some seconds in error ; and if the ship is in port 

 only a brief time, and the chronometer rating should occur 

 during a week when no observations can be made at the 

 observatory, the probability is that a very erroneous rate would 

 be assigned to the chronometer, for use during the coming 

 voyage. 



3. In the chronometer-shops the electric time-signals are 

 received from the observatory or standard clock at regular 

 intervals, usually every second or two seconds or minute. The 

 signals are received daily, and comparisons can be made with- 

 out regard to the weather, so far as distinguishing the signal 

 is concerned. The chronometer comparisons are made by 

 means of a ' 'hack' ' chronometer, which has been accurately 

 compared with the time-signal, and are made by some one per- 

 son. A skilled man can compare two chronometers without 

 having an error of more than one-tenth of a second, and the 

 comparison can be made several times a day if it is desirable; 

 and this is an important matter if the ship is to be in port 

 only a few days. During cloudy weather, when it is impossible 

 to make astronomical observations, and it sometimes happens 

 that a whole week will pass without an observation being 

 made, it is possible for the chronometer-maker, who usually 

 has on hand a large number of first-class time-keepers, to 

 keep his own standard clock nearer to the true time than that 

 given by the observatory clock, because he relies on the 

 average results given by a large number of time -pieces, while 

 an observatory seldom has more than two. In this case the 

 chronometer-shop rating of a ship's chronometer is much more 

 accui-ate than that by means of a time-ball. Accidental errors 

 in the standard clock-signals due to occasional mistakes made 

 by the astronomer can be detected (if they are large) by the 



