( 603 ) 
circles of Paris, Greenwich and Washington and those at the Alta- 
zimuth of Greenwich have been used. 
It is true that, in a remarkable paper, Znwestigation of corrections 
to Hansen’s Tables of the moon, with tables for their application, 
Newcoms brought together the corrections to be applied to the for- 
mulae by means of which Hansen calculated his tables of the moon. 
The paper contains the terms which had to be added according to 
the state of science in that year, and also an empirical correction 
determined by the most recent observations. 
Moreover a table of corrections for 1874 was given, founded on 
these data. But after having made a diagram representing, both the 
corrections found by direct observation and those furnished by 
Newcomp’s table, I came to the conclusion that the former was to 
be preferred *). 
As for the longitude of St. Denis, I will remark, that it has been 
determined by the French naval officer Germain in 1867 and 1868 
by means of 13 culminations of the first and 12 of the second limb. 
In the Connaissance des Temps of 1871 a short report of that deter- 
mination is to be found. Though the 25 results there given, agree 
tolerably well, this kind of determinations is always liable to the 
drawback that the difference in the constant error, made in observing 
the culmination of the moon’s limb and of the comparison stars, 
enters into the result, about thirty times magnified. There is no fear 
of such an injurious influence in a determination of longitude by 
occultations *). If the voyage to reach the isle of Réunion did not 
last so long, and if the Indian Ocean were not so wild and bois- 
1) The present state of science requires a correction of one of the tables of 
Newcoms. He points out (page 9) that the parallactic equation of Hansen is founded 
on the value 8”.916 of the solar parallax, whereas the value which he derived in 
1867 from all the available materials is but S'.S4S, which is less by 0.068. 
Further that later determinations require rather a diminution than an increase of 
that number. At present 8.800 is generally adopted as being the most probable 
value of the solar parallax, which is less than Newcomp’s value by 0'.648. The 
parallactie correction of Newcoms must therefore be increased 1,7 fold; in other 
words: three terms have to be added, viz. : 
+ 0".67 sin D + 0".05 sin (Dg) — 07.09 sin (D + 9’), 
where D represents the mean elongation of the moon from the sun, g the mean 
anomaly of the moon and g’ that of the sun. 
*) Newcomp says at the beginning of his paper above mentioned: “Determi- 
“nations of longitude from moon occultations are found by experience to be 
“subject to constant errors which it is difficult to determine and allow for. It 
“was therefore a part of the policy of the American Commission to depend on 
“occultations rather than upon culminations for the determinations of longitudes, etc.” 
