1H28 



To tliis same scale were coiupared a iniuiher of stars in the N. 

 and S.-strip on the Halbgitler-plates T, and 6V Here the lesult was: 

 S.-strlp: ordinai-v image 6\ — weakened image C\ = 



= 13.78— 1J,63 = 2,15 (75) 



N. -strip: ordinary image C\ — weakened image C\ ^ 



= 13,78— IJ, 48 = 2,30 (38) 



This gives for the absor|)tion of the Haibgitter in unities of the 

 provisional scale 2.22. In absolnte scale according to the statement 

 of Prof. Hkktzsprlnü at Potsdam there was fonnd for this absorption 

 1,9H3 nuign. All the intervals dednced here must therefore be 

 miilti[)lied Uy the factor 0,884, in order to express them in magni- 

 tudes (this means that a 10 times larger exposure gives a gain of 

 1,77 magn.). Then they are: 



B - A, = 0"',84; A, - J, = 0"',9r, ; .4, — .4, — 0'",92 ; 

 A, — .4, ^ 1"',02 ; A, - .1., = (i"',9(). 



In oi-dei- to express also the magnitudes themselves in absolnte 

 scale, 16 of the most brilliant slurs weir used, which aie contained 

 in the "Göltingei' Aktinometrie" ; from the magnitude of theii- 5'"' 

 and t)''' image was found : 



m - 11,55 0,884 (prov. m — 11,55). 



3. The Ihniting magnitiKk. The difference in limiting magnitude 

 will be equal to the differences in magnitude found here for the 

 same stars at various e\|)0sures, provided the conditions under which 

 the observations are made be absolutely identical. On the plates R 

 and .1, each star presents two ecpial images; all the double images 

 therefore that are at all discerinble are counted. With regard to the 

 exposures .4., A^ A^ and .4, on the other hand, a faint, scarcel\ dis- 

 tinguishable image must be looked for, iii a given spot by the side 

 of brighter images. If the chance that by the thu-tuations in the 

 conditions a star-image near the limit of visibility can be just 

 discerned =: (/, then the chance that two ecjual images are both 

 visible = c//*; in this case therefore more stars i-emain inxisible. With 

 such counting as on B and A^ therefore fewer will be counted, syste- 

 matically, than with the method employed for A., etc. For the 

 difference in limiting magnitude .4i— .4, the difference in magnitude 

 fonnd above can therefore not be used. 



In order to find this difference during the counting of the plates 

 A charts had already been drawn of those squares, where later on 

 the magnitude of all clearly visible stars was to be ascertained, on 

 which charts were indicated all the stars showing 2, 3, 4, 5 and 6 

 images. We must now find what magnitude, measured on B, forms 



