( 319 ) 



apsides. He further introduced two unknown quantities, viz. tiie 

 constant errors which might vitiate the observations of the two 

 observers Gill and Finlay. He thus also obtained a total of 29 

 unknown quantities. It need not be said that the sohition of about 

 400 equations with so many unknown quantities, is an enormous 

 labour. Still, owing to the help of some other computers, this labour 

 has been brought to a happy issue. 



We must not enter here into further particulars about this impor- 

 tant work, though we did not feel justified in omitting to mention it 

 altogether, 1 will only remark that it is not sufficient to determine 

 the position of the planes of the orbits of the satellites for one 

 epoch ; for as was already remarked the position of these planes 

 changes continually. It seems that these changes may be sufficiently 

 represented by assuming a regular retrogradation of the line of 

 intersection with a fixed plane, the inclination remaining the same. 



The main cause of this retrogradation is the polar compression of 

 Jupiter. It is desirable however to establish the amount of this 

 retrogradation by the observations, and to derive afterwards the 

 compression by means of this amount. Consequently the position of 

 the planes of the orbits has to be determined for different epochs. 

 In this respect too Mr. de Sitter has done good work, vide the 

 communication already mentioned, presented in the meeting of last 

 March by Messrs Kapteyn and E. F. van de Sande Bakhuizen. 



(8) The same volume, which contains the ecliptic tables of Damoiseau, 

 contains also in a second part (not mentioned on the title) tables 

 ''pour trouver les configurations des satellites de Jupiter." 



We have contemplated whether it would not be desirable not to 

 use these tables, unmodified, for our computations. We have therefore 

 taken note of the investigations of Souillart, Adams, Marth, Gill, 

 Finlay, and de Sitter, but it appeared that such a- course would 

 aggravate our labour very considerably. We would have had to 

 determine new elements for all the satellites and to compute new 

 tables. This would have caused considerable retardation, unnecessary 

 for our purpose, which was no other than to prepare astronomers 

 for the observation of the conjunctions visible in 1908. 



We therefore have based our computations on the tables of Damoiseau, 

 but we have first examined in how far they represent the observed 

 conjunctions. The following summary shows not only the difference 

 between the observation and the tables in the elongations x and x\ 

 of the two satellites, expressed in radii of Jupiter, but also their 

 difference in time. 



