825 
With this value of dn’ the computation was then carried out in 
exactly the same way as with dn. Notwithstanding the considerable 
difference between the functions /,’ and J, the general character of 
the results of the two computations is the same. 
The non-periodie part of the perturbation in longitude derived from 
the new computation is given in Table IV, which is entirely similar 
to Table I of Part I. We now find A,v' = — 230, Al’ = + 2939. 
Neglecting the term — Sp’ A,r’ and causing the perturbation to 
vanish for 1721 and 1865 by an appropriate choice of the constants 
of integration, we find the values given under the heading 4. If the 
term containing A,» is added, we get the value à',. The general 
TABLE IV. 
‚Year Saros An! | Aar eg a Ara? hel er si 
| | | 
1703.0 | | | | di. — 288 
| EN or Bae | 1918 | | 
| 1721.0 | | 0 | 0 
| Me 10442818 156 | — 621 | | 
‚17391 | | | — 490 | + 315 
| III | —21.9 | +2346 | — 550 | — 593 | 
| 17571 | | | = RISO: 272 
| IV | + 2.8 | +3300 | +334 | + 361 | 
1775.1 | | | | 224822) | 1/408 | 
| V | —12.3 | +3141 | —227 | + 202 | | 
| 17931 | | | | —1361 | + 476 | 
| Wich = 2 FP esRo.| is |e aa | 
| 1811.2 | | | 1388 | 4 337 
| VII | — 1.0 | +3466 | +193 | + 527 | 
1829.2 | | | —1444 | — 64 
| VIII | +14.9 | 43896 | +751 | + 957 
1847.2 | | | | ee aT eee 
IX | — 1.9 | +3452 | +160 | + 513 | | | 
1865.3 | | k | | NE ROE 
Relea. |e Sia. 190 | 21 1817 | 
1883.3 | | || + 699 | — 336 
| X1-| —19.6 | +2404 | —498 | — 445 | | | 
1901.3 | | | + 645 | —1655 
XH | 9.2°|:+ 2680.) + 572 | — 309°| | 
| 1919.4 | | | | + 229 | —3566 
| | | | 
character of the perturbation is very similar to that of the first com- 
putation. But the correspondence with the “great fluctuation’, which 
was apparent in the first computation, does not exist here. 
In the periodic part the agreement between the results of the two 
computations is even more complete. 
With reference to the reliability of these results it must be remarked 
that the function ./,’ has a wider range of variation depending on 
T, than J,, and consequently the possible error arising from the 
fact that 7, is only known to whole minutes is in the second com- 
putation much larger than in the first. Accordingly we find that the 
