474 I'. A. oYfciN [1914 



of radioaclive maller in ils ouler shell iL. c. pag. 204). Men 

 (iesuaglel ullaler Bkckkh ganske likelil: In Ihe present slale 

 of knowledge, esliniales of Ihe age of minerals founded on ra- 

 dioaclivitv can nol command conlidence iHull. (ieol. Soc. Ame- 

 rica, Vol.' 26, 191-5, j)ag. 195). 



JoLY kom i »An eslimate of Ihe geologicai age of Ihe earlh« 

 li! del resullal, al we Ihink Ihal il is al least jusliliahle lo 

 claim Ihal our present knowledge of solvent denudalion of Ihe 

 earlh s surface points lo a period of belween eighly and ninety 

 millions of years having elapsed since water condensed upon 

 the earlh, and rain and rivers and Ihe actions continually pro- 

 gressing in the soils hegan lo suj)ply Ihe ocean with malerials 

 dissolved from the rocks (Smithsonian Report for 1899, pag. 

 287), men han regner da samlidig op ikke mindre end syv fak- 

 torer, der vil virke til at gjøre delle overslag til et minimum, og syv 

 andre faktorer, der vil virke i motsat retning, nemlig til al gjøre 

 del samme oveislag lil el maximum (L. c. pag. 288\ Ved sedimen- 

 lalionen Ihe age is l'ound lo be 87 millions of years (Joly: 

 The Birlh Time of the World, 1915, pag. 12). Men vi kan her si 

 med P. Mc. Connell: »As the age of Ihe earth does nol parti 

 cularly concern the objecl of Ihis book, we may leave il now 

 and lel the scienlisls interested in the maller iight it oul among 

 themselves (The Elements of Agriculhiral (ieology, London 1902, 

 ])ag. IH). Men det har dog sin betydning al erindre derom, før 

 vi gaar over lil del emne, som her har en mere umiddelbar 

 interesse, nemlig de tidsforhold, der staar i forbindelse med den 

 kvartære lid og de forskjellige faser af denne. 



Førend vi gaar over hertil skal vi imidlertid erindre el ganske 

 almindeligt malhemalisk problem: The crilerion for Ihe besl 

 value is by no means evident. Perhaps it will be generall}' con 

 sidered thai the besl value is Ihe probable value, that is, Ihe 

 value thai is just as likely lo be exceeded as nol in the long 

 run, so thai the probability thai the observed value shall fall 

 shorl of Ihe probable value is \ 2 and the probability that the 

 observed value shall exceed the probable value is 1 2. Cerlainly 

 Ihe besl value cannol be delined as the most probable value, 

 that is, the value whose |)robabilily is a maximum, because il 

 turns oul that the probability is somelimes a maximum for more 

 than one value; the probability may even have equal maximum 

 for Iwo or more valnes^ (\V. K. Stohy: A new general Theory 

 of Errors, 1904, pag. 1()9). 



Samlidig med al Heulky s skildring af »The Paleogeographi 

 cal Relations of Antarctica opruller for os el konkret forelig 

 gende billede, vil man i de i vedføiede slutningsord fra hans 

 interessante afhandliui* foruten el saadanl oiisaa kunne se el 



