]\gt ; On of Mercury .
The writer wishes to tender his thanks to Mr. J. M. C. Paton , of the firm of NIanlove , Alliott , Ltd. , for much assistance , freely riven , in the making of also to Dr. R. T. Glazebrook , for giving hinl every facilit in tests made on his machine at the LaborRtor ; to the Royal Society for rant which defrayed some of the of this vork .
On Absolnte of By H. L. LLENI ) , LL. D. , , and H. llfoss , Demonstrator of Physics , College of Science and ( Received November 19 , 1910 , \mdash ; Read January 11911 .
) act .
's experiments on the absolute expansion of mercury were reduced by Wiillner , and uently b ] .
Their reductions differed by nearly 1 per cent. at C. 's own leductions differed by nearly 1 per cent. from either C. } ) ' later determinations by the thellnometer metbod , which were not absolute , reed fairly well with Wullner 's reduction of egnault at low temperatures , but differed from Wiillner in the opposite direction to Brooch by more per cent. when extrapolated to C. The ject of the present ation to repeat 's method on a scale with model.n appliances , and the apparatus was designed to secure an order of ccuracy of 1 in 10,000 , or C. , which , it is believed , has been substantially attained .
The principal modifications made in Regnault 's apparatus were as : In place of the single pair of hot and cold colunlls , each metres , employed by Reguault , six airs of hot and cold colunns .
each nearly 2 metres long , were connected in series as a multiple manometer , giving nearly eight times the ) ansion oinable with ) ( 2 ) The difference of level to be nleasnred directly referred to a standard Invar scale divided in millimetres divisions 4 lnicrons wide , by means of a telescopes with lliclometer to cm .
clilference of level .
to the fundamental to .
was cm .
, the limit of accuracy of reading was 1 in 20,000 of this interval , or C. } of the hot and cold columns , each nearly 2 metres , could be read to cm .
, the same order of accuracy .
The mean temperatures of the hot an cold On the of Mercury .
columns were observed by means of a pair of platinum thermometers equal in length to the columns , and reading to C. The possible error of this reading would not amount to more than C. at C. , and would be much less at lower temperatures .
( 4 ) Uniformity ) constancy of temperature were secured by a continuous circulation of oil , heated by means of resistance coils traversed by an electric current .
The method of electric heating permitted the most delicate regulation of the temperature , even at 30 C. , with the least possible disturbance of the surrounding conditions .
( 5 ) 's were taken with the cold column at the atmospheric temperature of to C. , which left the reduction to C. somewhat uncertain , and was the chief cause of the discrepancy in the reduction of his obseryations at low temperatures .
In the present investithis uncertainty was avoided by taking a special series of observations with the cold column at C. and at C. ( 6 ) The chief cause of the discrepancy between the reductions of Wullner and Brooch at temperatures appears to be that Brooch applied a correction for the conduction of heat along the cross-tubes connecting the hot and cold columns in Regnault 's apparatus .
These cross-tubes were not quite horizontal , and introduced some uncertainty in the effective of the columns .
In the present investigation , the cross-tubes were made of steel tube 1 mm. in bore , to diminish conduction of heat at the point where they emerged from the hot bath , and were held rigidly horizontal , by means of a specially desi , bracket , for the short in which the temperature changed from hot to cold .
The effective height of the hot column was thus rendered definite and accurately measurable .
With the apparatus above described , 94 complete sets of observations were taken on 33 days in the course of 1908 , 1909 , and 1910 , for 16 different ranges of temperature covering the interval from C. to C. The mean results at each temperature agree to 1 in 10,000 above 10 C. , or to C. below 10 C. , with the following formula for the mean coefficient from C. to C. The results cannot be represented satisfactorily by a linear formula for the mean coefficient over the whole range ; but for approximate work the following simple formula may be sufficiently exact to be of This formula gives results which are practically correct at 10 and 20 C. , with a maximum error of C. between C. and 20 C. , but the value of the mean coefficient is more than 1 in 400 too small at 30 C. Proof of the Theorem of Double Six of Straight .
597 The results of Eumorfopoulos*for the boiling-point of sulphur on the scale of the constant pressure air thermometer were about 1o C. lower than the previously accepted value C. , when reduced by the BrochRegnault formula for the expansion of mercury ; but are brought into practically perfect agreement with the old value , when reduced by reference to the results of the present investigation .
A Proof of the } of a Double Six of By H. , Sc. D. , ( Received Noyember 21 , \mdash ; Read November 24 , 1910 .
) * We assume that if two quadric surfaces have common two intersecting generators , their remaining common points lie upon a plane .
This is capable of sirnple geometrical proof .
Further , we prove a subsidiary theorem regarding eight lines , which we , 2 , 4 , , which satisfy certain conditions .
These conditions } first that in the schelne 1 2 o ) 4 1 ' 2 ' each line intersects the three which are not in the same or column of the as itself ( so that the lines can be arranged as two skew quadrilaterals , with the property each side of either intersects a particular side of the other and second that one of the two lines which call be drawn to meet 1 ' , 2 ' , 3 ' , 4'\mdash ; say ( \mdash ; intersects one of the lines which can ) drawn to meet 1 , 2 , 4\mdash ; say .
The conditions for the eight lines are then briefly expressible as the existence of a schelne 1 ' 2 ' 3 ' 4 ' 'Roy .
Soc. Proc , 1908 , vol. 81 , p. 339 .
If and be the common gene1atol s , intersecting in , a variable plane through will touch the quad1 respectively , say , in and , on , and the nges will be projective with the pellcil ; these have the ummon point , and will thus have another commou ) , say , lying 011 , let be the ) of ? ?
other than , at the qnad1ics touch .
If now be any point of the two quadrics not lying on or the plano A cuts the surfaces ely in conies having comlilon and , at A and B. These comics thus coincide in one , which } c.onstitute the remaining of intersection of the quadrics .

