364 Transactions of the Boyal Society of South Africa. 
meet the case. But if we are dealing with conditions indicating wet 
weather we are justified in taking any fall at all — even 0"0l in. — as 
showing a verification of the forecast. If, however, we are dealing with 
conditions indicating dry weather, we shall be quite in order if we take all 
falls between 0-01 in. and 0*05 in. as of no importance. The atmospheric 
conditions governing rainfall are so very erratic that we will be justified 
in this course of action ; for instance, suppose the indications are for rain 
and only 0"01 in. falls, the slightest change in the atmosphere, all indica- 
tions of which are present, will give a greater fall — it is a mere accident 
that the fall is so slight ; but if the indications are for dry weather and a 
fall of 0 05 in. follows, this also is accidental, and any atmospheric diffe- 
rence will make for less fall or none at all. In some cases when no 
rainfall is recorded at the Royal Observatory, it is seen to have fallen 
at L'Agulhas; this might be considered if the conditions are for rain, as 
it is not due to accidental circumstances ; but if the reverse be the case 
and still rain falls at L'Agulhas, then it may be due to causes further 
afield than those indicated by the triad of readings, and is thus outside 
our investigation. 
Taking, now, the barometric readings at the three stations, let us see 
how we can reduce them to a tabulated form. In the first place, we will 
divide them into primary groups : (I.) Where the general pressure is 
decreasing ; and (II.) Where it is increasing. In some cases a single 
station's barometer may have a reverse movement to the other two, but 
this will not upset our argument. 
In each of these two groups we can arrange for the following sub- 
groups : — 
(A) L'Agulhas most, Clanwilliam least. 
(B) L'Agulhas most, Eoyal Observatory least. 
(C) L'Agulhas most, other two equal. 
(D) Royal Observatory most, Clanwilliam least. 
(E) Royal Observatory most, L'Agulhas least. 
(F) Royal Observatory most, other two equal. 
(G) Clanwilliam most, L'Agulhas least. 
(H) Clanwilliam most. Royal Observatory least. 
(I) Clanwilliam most, other two equal. 
(J) Clanwilliam and Royal Observatory equally most. 
(K) Clanwilliam and L'Agulhas equally most. 
(L) Royal Observatory and L'Agulhas equally most. 
(M) All three moving simultaneously and equally. 
In one or two cases the whole of the readings remained steady for 
24 hours, but these I have not taken into consideration. It must be borne 
