XXI .
An Account of some Experiments on the Loss of Weight in Bodies on being melted or heated .
In a Letter from George Fordyce , M.D. F.R.S. to Sir Joseph Banks , Bart. P.R.S. Read April 28 , 1785 .
SIR , ALTHOUGH I have made many experiments on the subject of the loss of weight in bodies on being melted , or heated , I do not think it worth while to lay them all before the Society , as there has not appeared any circumstance of contradiction in them .
I shall content myself with relating the following one , which appears to me conclusive in determining the loss of weight in ice when thawed into water , and subject to the least fallacy of any I have hitherto made , in showing the loss of weight in ice on being heated .
The beam I made use of was so adjusted as that , with a weight between four and five ounces in each scale , 1/ 1600 part of a grain made a difference of one division on the index .
It was placed in a room , the heat of which was 37 degrees of FAHRENHEIT'S thermometer , between one and two in the afternoon , and left till the whole apparatus and the brass weights acquired the same temperature .
A glass globe , of three inches diameter nearly , with an indentation at the bottom , and a tube at the top , weighing about 451 grains , had about 1700 grains of New-River water poured into it , and was hermetically sealed , so that the whole , when perfectly clean , weighed 21503 1/ 3 2 of a grain exactly ; the heat being brought to 32 degrees , by placing it in a cooling mixture of salt and ice till it just began to freeze , and shaking the whole together .
After it was weighed it was again put into the freezing mixture , and let stand for about twenty minutes ; it was then taken out of the mixture ; part of the water was found to be frozen ; and it was carefully wiped , first with a dry linen cloth , and afterwards with dry washed leather ; and on putting it into the scale it was found to have gained about the 1/ 6 0 part of a grain .
This was repeated five times : at each time more of the water was frozen , and more weight gained .
In the mean time the heat of the room and apparatus had sunk to the freezing point .
When the whole was frozen , it was carefully wiped and weighed , and found to have gained 3/ 1 6 of a grain and four divisions of the index .
Upon standing in the scale for about a minute , I found it began to lose weight , on which I immediately took it out , and placed it at a distance from the beam .
I also immediately plunged a thermometer in the freezing mixture , and found the temperature 10 degrees ; and on putting the ball of the thermometer in the hollow at the bottom of the glass vessel , it showed 12 degrees .
I left the whole for half an hour , and found the thermometer , applied to the hollow of the glass , at 32~ .
Every thing now being at the same temperature , I weighed the glass containing the ice , after wiping it carefully , and found it had lost 1/ 8 and five divisions ; so that it weighed 1/ 1 6 , all but one division , more than when the water was fluid .
I now melted the ice , excepting a very small quantity , and left the glass vessel exposed to the air in the temperature of 32 degrees for a quarter of an hour ; the little bit of ice continued nearly the same .
I now weighed it , after carefully wiping the glass , and found it heavier than the water was at first one division of the beam .
Lastly , I took out the weights , and found the beam exactly balanced as before the experiment .
The acquisition of weight found on water 's being converted into ice , may arise from an increase of the attraction of gravitation of the matter of the water ; or from some substance imbibed through the glass , which is necessary to render the water solid .
Which of these positions is true may be determined , by forming a pendulum of water , and another of ice , of the same length , and in every other respect similar , and making them swing equal arcs .
If they mark equal times , then certainly there is some matter added to the water .
If the pendulum of ice is quicker in its vibrations , then the attraction of gravitation is increased .
For there is no position more certain , than that a single particle of inanimate matter is perfectly incapable of putting itself in motion , or bringing itself to rest ; and therefore , that a certain force applied to any mass of matter , so as to give it a certain velocity , will give half the quantity of matter double the velocity , and twice the quantity , half the velocity ; and , generally , a velocity exactly in the inverse proportion to the quantity of matter .
Now , if there be the same quantity of matter in water as there is in ice , and if the force of gravity in water be 1/ 2 8000 part less than in ice , and the pendulum of ice swing seconds , the pendulum of water will lose 1/ 2 8000 of a second in each vibration , or one second in 28000 , which is almost three seconds a day , a quantity easily measured .
I shall just take notice of an opinion which has been adopted by some , that there is matter absolutely light , or which repels instead of attracting other matter .
I confess this appears absurd to me ; but the following experiment would prove or disprove it .
Supposing , for instance , that heat was a body , and absolutely light , and that ice gained weight by losing heat ; then a pendulum of ice would swing through the same arc in 1/ 1 4000 less time than a similar pendulum of water ; for the same power would not only act upon a less quantity of matter , but a counter-acting force would also be taken away .
Till the experiment of the pendulum can be made , or some other equally certain be suggested and made , it would be wasting time to enter into conjecture about the cause of the gain of weight in the conversion of water into ice in a glass vessel hermetically sealed .
I shall only observe , that heat certainly diminishes the attractions of cohesion , chemistry , magnetism , and electricity ; and if it should also turn out , that it diminishes the attraction of gravitation , I should not hesitate to consider heat as the quality of diminution of attraction , which would in that case account for all its effects .
We come , in the next place , to take notice of the second part of the experiment , viz. that the ice gained an eighth part of a grain on being cooled to 12 degrees of FAHRENHEIT'S thermometer .
In this case , a variation may arise from the contraction of the glass vessel , and consequent increase of specific gravity in proportion to the air .
But it is unnecessary to observe , that this would be so very small a quantity as not to be observable upon a beam adjusted only to the degree of sensibility with which this experiment was tried .
In the second place , the air cooled by the ice above the scale becoming heavier than the surrounding atmosphere , would press upon the scale downward with the whole force of the difference .
If a little more than half a pint of air was cooled over the scale to the heat of the ice and glass containing it , that is , twenty degrees below the freezing point , the difference , according to General ROY'S table , would have been the eighth part of a grain , which was the weight acquired ; but the air within half an inch of the glass vessel being only one degree below the freezing point , I cannot conceive , that even an eighth part of a pint of air could be cooled over the scale to twenty degrees below the freezing point ; nor that the whole difference of the weight of the air over the scale could ever amount to the 32d of a grain .
I have , however , contrived an apparatus which is executing , in which this cause of fallacy will be totally removed .
I shall , therefore , rest at present the state of this part of the subject ; and leave it only proved , that water gains weight on being frozen .
I am , &amp;c.

