92 E. F. J. Love: 



value of g tor Sydney is 979 -687x111 sec-2 on the Potsdam system; 

 consequently the corresponding value for Melbourne is 979-977.cm 

 sec-2. This figure replaces both the previous incorrect one, viz.: 

 S79-969, and Borrass's semi-conjectural emendation of it, 979 993, 

 which is given in many tables. 



§ 3. Gravity at the Melbourne and Sydney Observatories 

 on the Potsdam System. 



This problem has already been discussed in part by Borrass,^ 

 but his discussion requires revision in view of Alessio's subsequently 

 published work,« and of the results in §2 above. 



Alessio's outfit was, in its main features, a replica of Hecker's;^ 

 their observations are characterised by much the same care and 

 attention to detail, and each determined the flexure correction at 

 every station, instead of trusting to its constancy as all previous 

 experimenters had done. They differ, however, in that Hecker used 

 five pendulums as against Alessio's four. They differ also in their 

 manner of observing, in that Hecker used his own clock, while Alessio 

 used clocks in regular use at the observatories, in preference to his 

 own, arguing quite justly that a clock in steady work is less liable 

 to systematic acceleration of rate than one recently set up;io on the 

 whole, Alessio's procedure seems to be slightly the more advantageous. 

 Comparison of their work discloses no other material advantage on 

 either side as regards method. Nevertheless, the mean errors of their 

 results differ, Hecker's being decidedly the smaller, especially for 

 Sydney. Alessio's Melbourne determination is therefore assigned 

 three-fourths, his Sydney determination one-half the weight of 

 Hecker's. 



For the relative weights, as compared with Hecker's, of other 

 determinations in which half-seconds pendulums were used, Borrass's 

 (I.e.) estimates are accepted. 



As regards the Kater pendulum determinations, Helmert (I.e.) 

 has assigned to that obtained at Sydney the same mean error, iOOlO, 

 as to those of von Elblein and Budik; the corresponding mean error 

 for the Melbourne determination — in which twice as many experi- 

 mental stations are involved — would be iO-014; but, for reasons given 

 in Appendix 1, this is increased to i 020; hence the Melbourne deter- 

 mination is allowed half the weight of the Sydney one. 



The data for the evaluation of g for the Melbourne observatory 

 are given in Table I. 



7. Assoc, geodes. int., eompt. rend. 16 ieme conf. gen., IlI.e vol., p. 224. 

 Frequent reference is made to this paper. 



8. Osservazioni Gravimetrische, Geneva, 1912. I owe my copy of this 

 paper to Dr. Baldwin's kindness. 



9. Hecker's masterly pendulum work is detailed in a series of mono- 

 .graphs published by " Zentralbureau der Internationalen Erdmessung," and 

 -' Koniglich-.preussisches geodatisches Institut." 



10. P\irther details on this point are given in Appendix 2. 



