546 
NATURE 
| APRIL 4, 1907 
ee 
of the eastern counties was chosen for the purpose. The 
group includes the county with the largest acreage under 
each of the ten crops named, with the single exception of 
grass. : 
The results for wheat are of especial interest in con- 
nection with Dr. Shaw’s conclusion as to the great import- 
ance of the autumn rainfall. Mr. Hooker confirms this, 
and finds, further, that the autumn is more important 
than any other period. The critical period is, however, 
probably somewhat shorter, the correlation of the produce 
with rain exhibiting a marked negative maximum for the 
thirty-seventh to forty-fourth weeks, the actual coefficient 
being —o-62; the coefficient with the rainfall of the cereal 
year as a whole is slightly greater still, viz. —0-69. 
There are two marked coefficients with the weather of the 
preceding summer, t.e. the summer of the year in which 
the seed for the crop was grown, viz. —o-49 with rain 
during the twenty-first to twenty-eighth weeks, and +0-51 
with temperature for the twenty-ninth to thirty-sixth 
weeks, indicating absence of rain during the flowering 
period and warmth at harvest as necessary for good seed. 
For barley the chief requirement appears to be a cool 
summer, and for oats the same thing holds, but the 
latter crop also demands rain in spring, as indicated by a 
coefficient of +0-70. In the case of turnips, the highest 
coefficient, +-0-55, is with the rainfall in June-July, i.e. 
the sowing season, this being partly due, in all probability, 
to the fact that in a dry season the turnip-fly will eat off 
a young crop almost as soon as it shows above the ground. 
In spite of prevalent opinion, there does not seem to be 
any need for rain in late summer. In the case of the hay 
crops, the great value of the rainfall in spring and early 
summer is very well brought out, the coefficients attain- 
ing sharply marked maximum values of more than o-7 in 
the spring. 
One conclusion of remarkable generality is reached, viz. 
the advantage of cool weather during the late spring and 
summer for all the crops dealt with (except, perhaps, pota- 
toes). Taking the period between the ninth and twenty- 
eighth weeks of the year, all the four coefficients with 
temperature are negative in the case of barley, oats, 
turnips, mangolds, and hay; for wheat and for beans three 
of the four coefficients are negative. The correlation is 
with cool weather as such, and not with rain, as the 
effect of rain is practically eliminated by the method used. 
The result seems to indicate that grain and roots yield the 
most bulky crops if developed gradually and equably; 
neither rains nor heat, in fact, seem to be good for the 
crop for some time before harvest. 
The paper also brings out very clearly another fact, viz. 
that the condition of the seed sown may be as important 
as the subsequent weather. As the condition of the seed 
is itself dependent on the weather of the year during which 
it was grown, this gives rise to the observed correlations 
between the crop and the weather of the seed year as well 
as that of the harvest-year. Further, the meteorological 
conditions necessary for seed quality appear to be, broadly 
speaking, somewhat opposed to those necessary for a bulky 
crop. Thus, in the case of wheat, absence of rain during 
the flowering period and warmth at harvest were found 
to be necessary for good seed, but for a bulky crop cool 
weather is desirable. Considering all the coefficients with 
temperature for the ninth to thirty-sixth weeks, for wheat 
only one out of six is positive in the harvest year, five in 
the seed year: for barley none is positive in the harvest 
vear, five in the seed year; for oats none in the harvest 
year, four in the seed year. This result would, by itself, 
suffice to account for the tendency observed in the case of 
cereals to an alternation of good and bad crops. 
Although there is considerable uncertainty in some of 
the less well-marked results owing to the small number 
of observations available (twenty-one years), the appli- 
cation of the laborious methods used appears to have fully 
justified itself by the conclusions which have been thereby 
reached. How great the labour must have been may be 
judged from the number of correlation coefficients— 
between six and seven hundred—which have been tabulated 
by the author. The paper is published, with an abstract 
of the discussion which took place at the meeting, in the 
Journal of the Royal Statistical Society for March. 
NO. 1953, VOL. 75] 
FLAME THE WORKING FLUID IN GAS AND 
PETROL ENGINES.* 
LAME produced by the combustion of inflammable gas 
or vapour and atmospheric air forms the working 
fluid of gas or petrol engines. 
Mechanical power can be obtained by means of flame in 
several different methods :— 
(1) By filling a vessel or cylinder with a mixture of gas 
and air, and igniting this mixture, a slight explosion is 
caused, and the excess pressure blows off through a valve. 
The temperature of the flame is very high, and so when 
it cools the pressure in the vessel is reduced below atmo- 
sphere. This reduction of pressure may be utilised by 
means of an engine operating by atmospheric pressure and 
discharging into a partly vacuous vessel, or by a piston 
moving into the vacuous vessel. This method may be 
called the explosion-vacuum method. 
A modification of this method exists which may be 
called the flame-vacuum method. In it the explosion is 
dispensed with. 
(2) By admitting a charge of atmospheric air and in- 
flammable gas or vapour at atmospheric pressure to a 
cylinder containing a piston, cutting off access to the atmo- 
sphere and the gas supply, and igniting the mixed charge, 
a mild explosion occurs; the pressure rises in the cylinder, 
and the piston is driven forward to the end of its stroke. 
(3) By supplying to a cylinder containing a piston a 
mixture of inflammable gas and air in a compressed state, 
and then igniting that mixture, a motive power can be 
obtained. 
These last two methods, (2) and (3), are respectively 
known as the non-compression method and the compression 
method of operation in gas and petrol engines. The two 
methods were illustrated by a_ specially constructed 
apparatus. In this apparatus the cylinder of a petrol 
engine was mounted so that the piston reciprocated 
vertically, and a guide rod was fixed vertically on the 
cylinder. A hundred-pound weight was arranged to slide 
on this guide rod, and arrangements made by which a 
given charge of gas could be introduced into the cylinder. 
It was also arranged that the weight could be let down on 
to the piston, firstly so as to rest without compressing the 
charge, and secondly allowing compression of about 10 Ib. 
per square inch. The mixture in the cylinder was ignited, 
and, in the case where the charge was not compressed, 
the weight was thrown up by the explosion and expansion 
a distance of about 10 inches. In the case where the 
charge was compressed, the weight was thrown up about 
18 inches, showing clearly the increased effect of the ex- 
plosion of a given charge when under compression. 
It is believed that this is the first time the effect of 
compression has been shown as a lecture experiment. 
(4) A cylinder is supplied with gas and air under 
pressure, but the mixture is ignited at a grating or shield 
as it enters the cylinder, and so the pressure in the cylinder 
never rises above the pressure at which it is supplied. 
The power here is obtained without any increase in 
pressure, and is due to the fact that a small volume of 
cool mixture, when inflamed, becomes a larger volume, 
so that although a pump may be used to compress mixture 
the expansion in the motor side is greater, although at 
the same pressure as the pressure in the pump. 
These four modes of action were all illustrated 
by means of specially constructed apparatus, in which 
the effect of the working flame could be seen. The 
four modes of action, and combinations or modifications of 
them, include all the fundamental methods used in obtain- 
ing motive power from flame which have been attempted 
by mankind for the last hundred years. In the year 1820 
the Rev. W. Cecil, of Cambridge, read a paper at the 
Cambridge Philosophical Society in which he described an 
engine which he had constructed to operate according to 
the explosion-vacuum method, and he states that at sixty 
revolutions per minute the explosions take place with 
perfect regularity. His engine consumed, he stated, 
17-6 cubic feet of hydrogen gas per hour. He also men- 
tions an engine operated in accordance with the second 
method, the non-compression explosion method, and one 
1 Abstract of a discourse delivered at the Royal Institution on Friday, 
February 22, by Mr. Dugald Clerk. 
