34 
maxima and good harvests frequently coinciding, although 
there is a tendency to excess of rain in such years. The excess 
comes in June and late autumn ; the latter after harvest is 
ended, the former when the crops are better for the rain. 
The years 1892-3, a period of maximum sun-spots, on the 
whole illustrate the general tendency. In 1892, August gave a 
rainfall slightly under the average, October one of the heaviest 
on record. This year, August must, locally, count as an 
exception. There was a slight excess, due to one or two 
heavy falls. At any rate the month proved to he one of the 
most glorious for harvest weather since the sixties. Were 
Ave, moreover, dealing with the district and not the city 
a considerable deficiency would have been shown. This 
is a capital instance of an accidental exception, such as 
makes it important not to theorise upon too narrow a range of 
observations. Last October gave, at York, a deficiency of 
8°/ 0 or 9°l 0 and the fall was below the average, also, over 
Yorkshire and adjacent parts. Yet here again there Avas 
considerable excess in other parts of the kingdom. There¬ 
fore, although the present maximum will doubtless show an 
excessive fall in October on the three-year mean, Ave must 
await next year’s results before Ave can tell Avhether August 
will result in a local deficiency. Whether the conditions 
existing here are true of the district can only be told by a 
similar examination in other localities. 
J. EDMUND CLARK. 
Note. —Since the above was written “ Knowledge , December , 
1893,” has published an interesting article by A. B. McDowall, 
M.A., upon the rain-curves since 1840, at eleven British stations, 
Bloxainised in five-year periods. The variation over the country 
is very striking. Sheffield (as might he expected) and Oxford 
give curves approaching nearest to that for York, but oftr high 
maximum for 1870 (1868-72) is prominent in neither, nor is it 
marked in any ea\~e Keswick and Rothesay, in both of which it is 
excelled by 1861, which at York was below the mean. All give 
a very marked minimum in ’55 or ’56, which hardly affects York. 
As to this paper we may observe :—(1.) Three years, rather than 
five Avould seem to be a preferable period to take; the latter 
including the vjliole sunspot cycle and so masking extremes. 
(2.) As noted above, there would seem to be great tendency to 
yearly values being masked by conflicting monthly variations. It 
Avould be very interesting if Mr. McDowall would develop these 
and so show which months in each case most affect the yearly 
resulis. Thus, the excess during the seventies is far more 
prominent away from the west coasts. This at York proved to 
be chiefly due to wet Septembers. 
