SOME PHYSIOLOGICAL RESULTS OF TEMPERATURE VARIATIONS. 125 
Col. I. Col. II. Col. III. Col. IV. 
September, = + 64 x 18 = + 115 subtracted from + 162 + A7 
October, = 0:0 - 0 y we 2) — 20 
November, = — Tl is — 128 fs fie si = * 84 
December, = — 87 Fs — 156 a » =— 148 — 8 
Col. I. = Temperature numbers treated as + and — quantities. 
Cole Il; = The above x. by 18. 
Col. III. = Conception curve treated as + and — quantities. 
Col. IV. = Resultant conception numbers after temperature numbers in Col. II. have been 
subtracted. 
When it is plotted as a curve it can better be studied (Chart III.). 
It is difficult to see the dependence of the resulting conception curve upon 
the above marriage curve, although in some particulars they do agree. 
It was only by taking into account other sets of data and other facts, that 
it was possible to prune from this conception curve parts which depended upon 
yet other factors, the resultant (completed in the above curve by the dotted 
lines) agrees most conclusively with the marriage curve. On examining the 
conception curve in this Chart, a large crest is seen in the months of June, July, 
August, September, and October,—broken, however, in August, another in 
March and April, and a third sharp peak in January. 
After careful comparison with curves produced from other sets of data, it 
was found possible to completely explain all discrepancies between this and the 
marriage curve. 
If the illegitimate births be examined (on which the marriage numbers can 
have no influence), a large maximum is always seen in October, corresponding 
to a maximum of conceptions in January. This results from the license which 
always exists in Scotland at the time of the New Year. That this license is 
great, every medical man will allow; and it is quite capable of producing a 
sharp rise of the conception curve. 
Now this maximum in January may be discarded, and this leaves us a curve 
with only two maxima, which can be much better compared with the marriage 
curve. 
There is yet a curious discrepancy in the curve, namely, the sudden dip of 
the curve in August, in the very middle, in fact, of a well pronounced crest. 
This really does not exist in the curve of a whole country ; but is probably 
due to a fluctuation—local in character—in the number of inhabitants. 
On looking over the numbers, not of the eight towns of Scotland alone, 
but of all Scotland, August was seen to give not a fall, but a rise (see dotted 
line in Chart TIL.). 
This dip in August does then not exist in Scotland, only in the towns. In 
all probability it is due to the outflux of inhabitants from the towns into the 
country at that season. 
