a blazing fire, we were delighted to make a change to the
comfortable dining-room, which communicated with the room we had just
left by means of folding doors, closed with the exception of just
sufficient space left at one end of the room to allow a waiter to pass
in and out. Very curiously, before the soup was finished, we became
aware that the candles which assisted the electric glow lamps (merely
for artistic effect) began to flare in a most uncandlelike manner--the
flames turning down, as if some one were blowing downward on the wicks;
and at the same time the complaints of "Draughts, horrid draughts!"
became general, and from every quarter. Finding that, as the dinner went
on, the discomfort became unbearable, even although the doors were shut
and screens put before them, I gave up dining, and took to scientific
discovery. The result of a few moments' observation induced me to order
"those gas jets," which I saw peeping out from among the foliage of the
electroliers, to be lighted up. In two or three minutes the flames of
the candles burned upright and steadily, and in less than ten minutes
the draughts were no longer felt; in fact, the room became really
comfortable.

The reason of the change was simple. The stratum of air lying up at the
ceiling was comparatively cold. The column of heated air from the bodies
of the twenty guests, joined to the heat produced by the movements of
themselves and the waiters, together with the steam from the viands and
respiration, displaced the colder air at the ceiling, and notably that
coldest air lying against the surface of the glass. This cold air simply
dropped straight down, after the manner of a douche, on candles and
heads below. The remedy I advised was the setting up of a current of
hotter steam and air from the gas burners, which stopped the cooling
effect of the glass, and created a stratum of heated steam and air in
slow movement all over the ceiling. The effect was a comfortable
sensation of warmth and entire absence of draught all round the table.
Later on, to avoid the possibility of overheating the room, the gas was
put out, and the electric lights left to themselves. But before we left,
the chilliness and draughts began to be again felt.

The incident here narrated occurred at the end of the month of April
last, when we might reasonably have hoped to have tolerably warm nights.
It is therefore clear that in this instance neither electricity nor
candles could eff