Epwin.—On the Principle of New Zealand Weather Forecast. 41 
exaggerate whatever was due to irregularity of the land surface; and it was 
only by closely watching the changes accompanying the passage of every 
low area, and patiently endeavouring to assimilate these conditions, that 
it became evident that the complications already alluded to were the result 
of there frequently being more than one low area in the vicinity of the 
country at the same time. 
One of the principal aids in arriving at this conclusion is dicen from the 
substitution of contour lines for the isobar as generally drawn; these isobaric 
contours are drawn in the direction which the wind blows to—flow in fact 
with the wind—and the idea of adopting this plan originated in an endeavour 
to comprehend local peculiarities, for it was evident after investigating such 
remarkable distributions of pressure as those already referred to, that, if 
warnings were sent of an approaching gale, they must make known the 
limits within which it would be locally experienced, for if the warning 
were based upon theoretical principles it would frequently be incorrect, and 
the calculation of gradients under these circumstances seemed rather 
difficult. 
In drawing these isobaric contours, the mobile nature of air has to be 
taken into consideration. We all know how susceptible it is of expansion 
or contraction, to changes of temperature and pressure; and this being the 
case, it is but a step further to allow that a cyclonic wind, whose shape 
may be assumed as circular while beyond the influence of land, may 
become much changed in shape when it approaches a mountainous country, 
and by reference to Piddington and other writers upon circular storms it 
will be found that this point has already been under notice; but a little 
further consideration will show that this is not the only difficulty that has 
to be dealt with, as it will soon become evident that the subject contains 
some complex features; for while the front or advancing curves of a cyclone, 
encountering high land, become retarded and deflected, the centre continues 
to press forward with undiminished speed, and consequently the isobars 
become packed in the vicinity of the retarding influence; but from these 
causes the curves will have a tendency to open out at the rear of the cyclone 
(Pl. I., fig. 2), and as the land has a retarding effect upon the cyclone’s 
advance, so will it also act upon it after the centre has passed, causing the 
rear curves to be extended; but the advancing curves will be compressed 
under either condition ( Pl. I., fig. 3). 
In drawing these isobaric contours, it must be constantly remembered 
that they must never cross each other, and that the object must be to endea- 
vour to depict a series of concentric rings more or less bent out of their true 
shape ; each tenth of an inch of barometric difference must be thas mec 
flowing in the direction toward which the wind moves, and each 
