jeg til < (bsefvationsuhr et Lomniechronometer. der var prø- 

 vet paa Observatoriet i Neufchatel. 



Boxchronometrene havde sin Plads ombord i et Skab 

 i Arbeidssalonen. De bleve liver Morgen optrukne og 

 sammenlignede indbyrdes. I L876 tjente Kullberg som 

 Hoveduhr og i 1877 og 1878 Reid. Kun Hovedubret har 

 været benyttet til Længdebestemmelserne, da det viste sig 

 at have en meget jevnere Gang entl de øvrige. Under 

 Rejserne blev det af de daglige Sammenligninger consta- 

 teret. at der ikke indtraf nogen merkelig Forrykkelse i 

 Høveduhrets daglige Gang. 



Ved Observationer paa Land eller paa Dæk benytte- 

 des saagodtsom uden Undtagelse et at de andre Boxchrono- 

 metre eller et Lommechrononieter, stundom ogsaa et al- 

 mindeligt Lommeuhr. I ethvert Tilfælde blev Obsérvations- 

 uhret sammenlignet med Hovedulnvt før' og etter ( Ibser- 

 vationerne. 



For Expeditionen tiltraadte sine Rejser, bleve Box- 

 chronometrene. med Undtagelse af Reid. daglig sammen- 

 lignede med Pendeluhret paa Observatoriet i Bergen af 

 dettes Bestyrer. Hr. A strand. Under Expeditionens Op- 

 hold i norske Havne blev deres Stand for Greenwich Mid- 

 deltid bestemt ved de fra Observatoriet i Christiania gjen- 

 nem Telegrafen givne Tidssignaler. Disse gives hver Søn- 

 dag og hver Onsdag Morgen. Signalapparatet (Morse's) 

 staar lige ved Siden af < (bservatoriets Normalpendel. Der 

 gives liver Gang 3 Signaler, nemlig 8'' 59™ 0», i 1 " 0™ 8 og 

 9* 1'" 0* Greenwich Middeltid om Søndagene .og 7' 59™ O, 

 s ; ()»> O og 8'' l m 8 om Onsdagene. Tidssignalet, der 

 høres meget skarpt paa Modtagelsesstationéns Morse-Appa- 

 rat. bestaar i et enkelt Slag. For at skille mellem de 3 

 Signaler slaaes efter det første 1 Dobbeltslag, etter det 

 andet 2 og efter det 3die Signal 3 Dobbeltslag. Signa- 

 lerne sendes til alle norske Telegrafstationer. 



Med Hensyn til Nøjagtigheden af de i det Følgende 

 givne Tids-. Længde og Bredde-Bestemmelser maa jeg be- 

 merke Følgende: 



Den af Beregningerne udledede sandsynlige Fejl af 

 en enkelt observeret Højde (paa Land med Stativ og Kvik- 

 sølvhorizont) er omkring + 5". Der er imidlertid, som 

 det paa sit Sted skal vises. Tegn til. at der. foruden de 

 egentlige tilfældige Observationstejl. optræder constante Fejl, 

 hvis Aarsag kunne ligge i forskjellige Omstændigheder, som 

 i Bestemmelsen af Indexfejlen, mangelfuld Justering af In- 

 strumentet. Excentricitet m. m. Evad Endexfejlén angaar, 

 da er den i Regelen bestemt samtidig med Observationerne 

 og ved gjennemsnitlig 4 Satser Dobbel^contacter af Sol- 

 renderne. Middelfejlen at Resultatet af en enkelt Sats 

 tinder jeg at være + 5". 7 og Middelfejlen for en Bestem- 

 melse af Indexfejlen skulde saaledes være + 2 // .85. In- 

 strumentet holdtes altid godt justeret og om Excentricitet 

 af nogen merkelig Virkning nævner Prof. Hansteen ikke 

 Noget. Ved Observationerne i Hammerfest (Fuglenes) og 

 i Bodo antydes imidlertid Tilstedeværelsen af constante Feil 

 i Højden af respektive -f 8" og - 8". En Del heraf kumle 



box-chronometer by Reid. and in the latter year I took 

 observations witb a pocket-chronometer, tested at the Ob- 

 servatory of Neufchatel. 



The box-chronometers were kept in a cupboard in the 

 work-room. They were wound up every morning and duly 

 coinpared. In 1876, Kullberg's served as chief time-keeper, 

 in 1867 and 1878 that by Reid. For determinations of 

 longitude, exclusive use was made of the chief timé-keeper, 

 its rate having proved much more uniform than that 

 of the other chronometers. On each cruise the result of 

 the daily comparison showed that no appreciable disturbance 

 had occurred in the diurnal ,rate of the chief time-keeper. 



For observations on shore, or from the deck of the 

 vessel, we used almost without exception one of the other 

 box-chronometers. or a pocket-clirouometer. nav sometimes 

 a watch. The chronometer or watch, whichever it might 

 be. was. however, tnvariably coinpared with the chief timé- 

 keeper botli before and after the observations. 



Previous to the departure of the Expedition on its 

 several cruises, the b<5x-chronometers. witb the exception 

 of Eteid's, were daily compared with the standard-clock of 

 the Bergen Observatory, by the Director, Mr. Astrand, 

 During the stav of the Expedition at Norwegian ports. 

 their error on Greenwich mean time was determined by 

 the time-signals telegraphed trom the Observatory at Chri- 

 stiania. These signals are sent every Sunday an,d Wed- 

 nesday morning. The signalling apparatus (Morse's) stands 

 close beside the standard-clock of the Observatory. Three 

 signals are given, viz. at 8* 59™ 8 , Q h U m 8 , and i»'' 1"' 8 , 

 Greenwich mean time, on Sundays, and at 7" 59™ IJ 8 , 8'' 

 0»' 0" and S' ]'» ir on Wednesdays. Each signal, distinctly. 

 delivered by the apparatus of the receiving-station. consists 

 of a single click. As a means of reådily distinguishing be- 

 tween the 3 signals, the first is followed by a double-click, 

 the second by '2 double-clicks, and the third by 3. These 

 signals are transmitted to all Norwegian telegraph-stations. 



As regards the accuracy of the observations for de- 

 termining time, latitude, and longitude. I must observe 

 as 'follows: — 



The computed probable error of a single observed 

 altitude (using on shore the sextant with a stand and mer-: 

 cury-horizon) is about + 5". . Meanwhile, there is reason 

 to believe, as will afterwards appear, that. apart from the 

 accidental errors of observation, certain constant errors 

 occur. arising probably from various sources, such as the 

 determination of the index-error, imperfect adjustment of the 

 instrument, excentricity, &c. With regard to the index- 

 error. this has usuallv been determined when taking the ob- 

 servations. and on an average from 4 sets of double con- 

 tacts of the solar limbs. The mean error of the result of 

 one double contact I found to be + 5". 7, and the mean 

 ! nor of the determination of the index-error should ac- 

 cordingl) have been + 2". 85. The instrument was always 

 kept well adjusted, and oi excentricity that could have any 

 appreciable effect, Hansteen says nothing whatever. Mean- 

 while, the observations tåken at Hammerfest (Fuglenes) and 

 at Bodo indicate the existence of constant errors. in the 



