88 Proceedings of the Royal Society of Victoria. 



of the balance spring would render it quite untrustworthy. 

 The observation and reduction of lunar distances has been 

 a favourite pastime with me for fort}^ years, and from the 

 experience thus obtained, I have formed such a high 

 estimate of the method, that I have viewed with sorrow its 

 gradual disuse. Many reasons have been given for this 

 state of affairs — one is, that seamen have not time for the 

 necessary calculations, but with Thomson's tables, which 

 are quite accurate enough for navigation, the time required 

 for a reduction is only ten minutes. Another, and this I 

 consider the true reason, is that the results could not be 

 trusted. In the consideration of this it may be stated that 

 about thirty years ago the lunar tables were so inaccurate 

 that an error of twent^^-two miles of longitude would 

 sometimes be due to the Nautical Almanac alone. In 18G2, 

 Hansen's tables were first used in this work ; these gave the 

 place of the moon very accurately for some years, but 

 commenced to diverge a little time ago, to counteract 

 which, the Nautical Almanac applies Newcomb's empirical 

 cori'ection to the places of the moon used in computing 

 the lunar distances, and the outstanding error is now very 

 small indeed. Besides the errors of the Nautical Almanac, 

 the)-e are those of the sextant, which owing to excessive 

 competition and lowering of price, are larger than formerly. 

 It is quite surprising how rai-ely one meets with a really 

 good sextant nowadaj's. Those made by the celebrated 

 Troughton at the beginning of this century are still 

 unsurpassed. I took the following excerpt from the 

 catalogue of a manufacturer who exhibited at our last 

 exhibition : — " We guarantee any of our best sextants to 

 have a small margin of error (under 3', and generally under 

 ]' of arc)." As the index error is easily and always 

 ascertained by the observer, I suppose the error referred to 

 is independent of this, principally excentric and division 

 error, which is so troublesome to measure that it is generally 

 neglected by seamen ; and how would these erj'ors affect a 

 longitude determined from a lunar distance ? by about 

 thirty times their amounts — 90' and 70' of arc respectively. 

 I once had brought under my notice a modern sextant, 

 where the excentric error amounted to 9'. If such a one 

 had been used in a lunar distance, the error of longitude 

 would have been about four degrees and a half. In other 

 cases, sextants originally very good, have been ruined by 

 incompetent repairers, or bungling treatment by their 



