508 MA.JOE-GENEEAL SABIXE ON THE EESEXTS OE HOHELT OBSEEYATIONS 
deflection at certain hours of the day may be easterly at one station whilst it is westerly 
at another, augments the probabihty, that, whenever the effects of the two classes of 
phenomena are separated so that the approximate laws of each may be known, the 
apparent anomalies will disappear, and the solar variation will be found, after the 
disturbance-variation has been eliminated, to corr’espond in its times of easterly and 
westerly elongatiorr with the order of the phenomena as generally observed in the same 
hemisphere : whilst on the other hand we may derive from the great dissimilarity in the 
laws of the disturbance- variation and those of the regular solar variation the salutari- 
caution, that until the disturbance- variation at any particular station has become known bi- 
actual iirvestigation, no certain inference can be dramr as to its direction or amount at 
diflererrt hours ; or as to the modification of the turning-hours which its junction with 
the solar-diurnal variation may produce. 
The subjoined Table (V.) exhibits in parallel columns the diurrral variation at Point 
Barrow as derived, — 1, in column A, from the whole of the observations including the 
disturbances ; it is here the joint eflect of both classes of phenomena ; and 2, in column B. 
from the observatiorrs when the disturbed observations have been separated and omitted. 
Oir comparirrg these two columns, it is seen that the time of westerly extreme, wlrich. 
when the whole of the observations are takerr into account, falls as late as 11 P.M., is 
restored iir column B, which represeirts approximately the regular solar-diimral variation, 
to the same hour at which it occurs at Tororrto, \iz. betweerr 1 and 2 p.ii. In column D 
of the same Table, the solar-diurrral variation at Toronto has been reproduced from the 
third volume of the ‘Toronto Observations’ (Table LXV. Colurrrrr 4, p. Ixxxriii}, for the 
purpose of facilitatirrg its comparison with the corresporrdirrg values iir columir B. In 
pirrsuirrg this comparison, we fiird that the characteristics of this variation at the t^'o 
stations are as nearly the same as can reasorrably be expected, considerirrg that it is only 
the larger disturbances that have been separated, and that sorrre minor efiects of the 
sanre class must still remain in the body of the observations after the rerrroval of the 
distm’bances of greatest magnitude. Thus we find at both statioirs that the north end of 
the magnet is at its easterly extreme about 7 or 8 a.m. ; that hr returnurg fr'om thence 
towards the west its motion is more rapid tharr at any other part of the twerrty-foiu- hours ; 
arrd that it passes through its mean dfrection about 11 A.M., arrd reaches its westeriy 
extreme a little after 1 p.m. Its subsequent motion, towards the east, is comparatively 
slower at Point Barrow tharr at Toronto, and is checked at both statiorrs by a small 
retrogression towards the west, which is greatly dimmished both hr corrtirruance and 
anrorrirt by the omission of the larger disturbances, markhrg thereby the source to which 
it is most probably due. The most irotable difiererrce which we find hr the comparison 
of columns B and D is, as afready noticed, that the retunr towards the east dmfrrg the 
hours of the afterrrooir arrd night takes place comparatively rrrore slowly at Pohrt Barrow 
than at Tororrto. At the latter station the magnet passes through its mean dh'ection 
between 7 and 8 p.m., wlrereas at Poirrt Barrow it does not do so urrtil 1 or 2 a.m. Now 
this is precisely what should take place on the supposition that a portioir of the easterly 
