MECHANIC S. 
liow much the mercury has been lower than it is; which 
being lubtrafted from the prefent height will give the 
lowelt point to which the mercury has fallen. That the 
thermometer may he fitted for a new obfervation, the 
mercury mutt be made to run back from the ball into the 
ihorter leg, by inclining the tube and heating the ball. 
In 178a Mr. Six propofed a felf-regifiering thermome¬ 
ter, which unites the properties of both the preceding, 
fliowing the greateft degree of heat and alfo that of cold 
upon one inftrument. It is properly a fpirit-of-wine ther¬ 
mometer though mercury is alfo employed for fupporting 
an index. This inftrument is reprefented at fig. 38. ab 
is a thin tube of glafs 16 inches long, and 5-i6ths of an 
inch caliber : c d e and f g h are fmaller tubes, about 
1 -20th of an inch caliber, which diameter however is en¬ 
larged to half an inch from h to z, which is two inches in 
length. Thefe three tubes are filled with highly-reftified 
fpirit of wine, except the fpace between d and g, which 
is filled with mercury. As the fpirit of wine contrails 
or expands in the middle tube, the mercury falls or rifes 
in the outlide tubes. An index, fuch as that reprefented 
at fig. 39. is placed on the furface, within each of thefe 
tubes, fo light as to float upon it. k is afmall glafs tube 
3 +ths of an inch long, hermetically fealecl at each end, 
and inclofing a piece of (teel wire nearly of its own length. 
At each end,/, m, of this fmall tube, a fhort tube of black 
glafs is fixed, of fuch a diameter as to pal's freely up and 
down within either of the outlide tubes of the thermo¬ 
meter ce or fk. From the upper end of the index is drawn 
a fpringof glafs to the finenefs of a hair, and about 5-7ths 
of an inch long; which being placed a little oblique, 
preffes lightly againft the inner furface of the tube, and 
prevents the index from defeending when the mercury 
defcends. Thefe indexes being inlerted one into each 
of the outfide tubes, it is eafy to underftand how they 
point out the greateft heat or cold that has happened in 
the obferver’s abfence. When the fpirit of wine in the 
middle tube expands, it prelies down the mercury in the 
tube c e, and raifes it in the tube kf ; confequently the 
index on the left-hand tube is left behind, and marks 
the greatelt cold, and the index in the right-hand tube 
rifes, and marks the greateft heat. But this thermometer 
is liable to au accident which, if not attended to, may very 
much impair the accuracy of its indications. A fmall 
portion of air is liberated from the alcohol it contains, 
which, getting into the tube, is for-nd fometimes to in- 
creafe, till it occupies as much as 5 0 in the fcale, making 
the refults fo much too high. This happens annually, at 
the firft approach of frofty weather. The remedy is to 
bring the inftrument to the fire, and caufe the bubble to 
paf6 to and fro in the warm fpirit, getting it, if pofiible, into 
the large tube ; by which means it is gradually abforbed, 
and does not foon re-appear. 
In the year 1790, a paper was prefented to the Royal 
Society of Edinburgh, describing two thermometers, new¬ 
ly invented, by Dr. John Rutherford of Middle Balilifti ; 
the one for regifteriug the higheft, and the other for regif- 
tering the loweft, degree of heat to which the thermome¬ 
ter has rifen or fallen during the abfence of the obferver. 
An account of them may be found in the third volume 
of the Tranfadtions of the Society. 
A new felf-regiftering thermometer has more lately 
been invented by Mr. Keith of Ravelftone, which we con- 
fider as the molt ingenious, fimple, and perfedi, of any 
which has hitherto appeared. Its fimplicity is fo great, 
that it requires only a very fiiort defeription to make it 
intelligible. It is conftituted, firft, of a thin glafs tube 
A B (fig. 4.0.) about fourteen inches long and fths of an 
inch caliber, clofe or hermetically fealed at top. To the 
lower end, which is open, there is joined the crooked glafs 
tube B E, feven inches long, and -^ths of an inch caliber, 
and open at top. The tube A B is rilled with the ftrongeft 
fpirit of wine, and the tube BE with mercury. This is 
properly a fpirit-of-wine thermometer; and the mercury 
is ufed merely to fupport a piece of ivory or glafs, to 
which is affixed a wire for railing one index or dc-preffing 
another, according as the mercury rifes or falls. E is a 
fmall conical piece of ivory or glafs, of fuch a weight as 
to float on the furface of the mercury. To the float is 
joined a wire called the Jloatwire, which reaches upward* 
to H, where it terminates in a knee bent at right angles. 
The float-wire, by means of an eye at a, moves eafily 
along the fmall harpflehord-wire G K. LL are two in¬ 
dexes made of thin black oiled filk, which Hide upwards 
or downwards with a force not more than two grains. 
The one placed above the knee points out the greateft 
rile, and the one placed below it points out the greateft 
fall, of the thermometer. 
When the inftrument is to be prepared for an obferva- 
tion, both indexes are to be brought clofe to the knee H. 
It is evident, that, when the mercury rifes, the float and 
float-wire, which can be moved with the fmaileft force, 
will be pulhed upwards till the mercury becomes ftation- 
ary. As the knee of the float-wire moves upwards, it will 
carry along with it the upper index L. When the mer¬ 
cury again fubfides, it leaves the index at the higheft 
point to which it was raifed, for it will not defeend by 
its own weight. As the mercury falls, the float-wire does 
the fame ; it therefore brings along with it the lower in¬ 
dex L, and continues to deprefs it till it again becomes 
ftationary or afeends in the tube; in which cafe it leaves 
the lower index behind it, as it had formerly left the up¬ 
per. The fcale to which the indexes point is placed pa¬ 
rallel to the flender harpfichord-wire. It may be feen 
more diftindily in fig. 41. That the fcale and indexes may 
not be injured by the wind and rain, a cylindrical glafs 
cover, clofe at top, and made fo as to exadtly fit the part 
DF, is placed over it. 
The ingenious inventor has another improvement itv 
contemplation, which, if upon trial it be found to anfwer, 
will make this thermometer as perfefl as can be defired, 
provided there do not arife fome errors from the variable 
preflure of the atmofphere. He propofes to adopt clock¬ 
work to the inftrument in fuch a way as to regifter with 
the utmoft precifion the degrees of heat and cold for every 
month, day, and minute, in the year. An account of 
this latter improvement may be feen in Nicholfon’s Jour¬ 
nal, vol. iii. 4to. or Edin. Tranf. vol. iv. 
The common contrivance for a felf-regiftering thermo¬ 
meter, now fold in 1110ft of the London fhops, con lifts 
Amply of two thermometers, one mercurial and the other 
of alcohol, as fhown at fig. 4a, having their Items hori¬ 
zontal. The former has for its index a fmall bit of mag- 
netical fteel wire, and the latter a minute thread of glafs, 
having its two ends formed into fmall knobs by fufion in 
the flame of a candle. The magnetical bit of wire lies in 
the vacant fpace of the mercurial thermometer, and is 
puflied forward by the mercury whenever the tempera¬ 
ture rifes, and puffies that fluid againft it; but, when the 
temperature falls and the fluid retires, this index is left 
behind, and confequently (hows the maximum. The 
other index, or bit of glafs, lies in the tube of the fpirit- 
thermometer immerfed in the alcohol; and, when the 
fpirit retires by depreffion of temperature, the index is 
carried along with it in apparent contadt with its interior 
furface ; but on the increafe of temperature the fpirit goes 
forward and leaves the index, which therefore ihows the 
minimum of temperature fince it was fet. As thefe in¬ 
dexes merely lie in the tubes, their refiftance to motion 
is altogether inconfiderable. The fteel index is brought 
to the mercury by applying a magnet on the outfide of 
the tube, and the other is duly placed at the end of the co¬ 
lumn of alcohol by inclining the whole inftrument. Mr. 
Nicholfon explains the operation of this inftrument thus : 
“ When the furface of the column of fpirit is viewed by 
a magnifier, it is feen to have the form of a concave he- 
mifphere, which fhows that the liquid isattradled by the 
glafs. The glafs in that place is confequently attrafled 
in the oppofite diredtion by a force equal to that which 
is fo employed in maintaining that concave figure j and, 
1 if 
