( 202 ) 



„article was written, but I interpohitc this statement with regard 

 „to it in order to enable astronomers to decide as to tlie justness 

 „of the views therein set forth. Both of the gentlemen of the Leiden 

 „observatory strenuously maintain that the mean period is more than 

 „431 days, and that it is invariable. The formula Y" (that is the 

 result given by E. F. v. d. S. B.) „is deduced by a peculiar and 

 „arbitrary treatment of the results of observation, its initial epoch 

 „being based on the Leyden observations alone, on the alleged 

 „ground that its errors are far smaller than those of all other series, 

 „which are rejected. T must however deny the propriety of assigning 

 „a weight of zero, relative to Leyden, to the extensive and precise 

 „series at Pulkowa between 1863 and 1882 with the Vertical Circle 

 „and Prime Vertical Transit." 



I shall now take the liberty to add on my part some remarks to 

 these opinions of Dr. Chandler. At the same time I shall make 

 use of the opportunity to consider the problem of the length of the 

 14-monthIy peiiod somewhat more closely, which consideration will 

 natui'ally lead to the discussion of the results on this point arrived 

 at by Chakdler in his last paper. 



2. In the first place in i egard to the grievances raised by 

 Chandler against my manner of treatment I will grant at once 

 that, by not using the results obtained at Pulkowa in the years 

 1863 to 1882, I would have committed a gross error, if it had been 

 my purpose to include in my investigations, in an independent way, 

 the observations before 1890. This however was in nowise the case. 

 It was simply my intention to submit to a discussion only those 

 obtained in the period 1890 to 1897; but as from these alone the 

 length of the 14-monthly period could naturally be derived with 

 but slight accuracy I had recourse to the results formerly deduced 

 and compiled by H. (i. v. D. Sande Bakhuyzen. It seemed unde- 

 sirable however to use all these results. In the first place, for 

 reasons to be stated hereafter, I thought it necessary to exclude 

 those of an epoch before 1860. Further consideration then led me 

 to restrict myself, in the deduction of a j^rocw/o?»?/ result, as far as 

 concerns the observations between 1860 — 1880, wholly to the Linden 

 results. 1 determined on this course because these proved to have much 

 smaller mean errors than all others of the same time, in as far as 

 they had been treated by H. G. v. D. S. Bakhuyzen, whilst moreover 

 these Leyden results proved to lie about midway between the others; 

 so that by including these the final result could not be modified 

 to any considerable amount. 



I might peihaps have pointed out still souunvliiit iiKirc clearly the 



