Being desirous of examining how the interior of a mass of ice is affected by a beam of radiant heat sent through it, I availed myself of the sunny weather of September and October 1857. The sunbeams, condensed by a lens, were sent in various directions through slabs of ice. The path of every beam was observed to be instantly studded with lustrous spots, which increased in magnitude and number as the action continued. On examining the spots more closely, they were found to be flattened spheroids, and around each of them the ice was so liquefied as to form a beautiful flower-shaped figure possessing six petals. From this number there was no deviation. At first the edges of the liquid leaves were unindented; but a continuance of the action usually caused the edges to become serrated like those of ferns. When the ice was caused to move across the beam, or when the beam was caused to traverse different portions of the ice in succession, the sudden generation and crowding together of these liquid flowers, with their central spots shining with more than metallic brilliancy, was exceedingly beautiful.

In almost all cases the flowers were formed in planes parallel to the surface of freezing; it mattered not whether the beam traversed the ice parallel to this surface or perpendicular to it. Some apparent exceptions to this rule were found, which will form the subject of future investigation.

The general appearance of the shining spots at the centres of the flowers was that of the bubbles of air entrapped in the ice; to examine whether they contained air or not, portions of ice containing them were immersed in warm water. When the ice surrounding the cavities had completely melted, the latter instantly collapsed, and no trace of air rose to the surface of the water. A vacuum, therefore, had been formed at the centre of each spot, due, doubtless, to the well-known fact that the volume of water in each flower was less than that of the ice, by the melting of which the flower was produced.

The associated air-and-water cells, found in such numbers in the ice of glaciers, and also observed in lake ice, were next examined. Two hypotheses have been started to account for these cells. One attributes them to the absorption of the sun’s heat by the air of the bubbles, and the consequent melting of the ice which surrounds them. The other hypothesis supposes that the liquid in the cells never has been frozen, but has continued in the liquid condition from the nÉvÉ or origin of the glacier downwards. Now if the water in the cells be due to the melting of the ice, the associated air must be rarefied, because the volume of the liquid is less than that of the ice which produced it; whereas if the air be simply that entrapped in the snow of the nÉvÉ, it will not be thus rarefied. Here, then, we have a test as to whether the water-cells have been produced by the melting of the ice.

Portions of ice containing these compound cells were immersed in hot water, the ice around the cavities being thus gradually melted away. When a liquid connexion was established between the bubble and the atmosphere, the former collapsed to a smaller bubble. In many cases the residual bubble did not reach the hundredth part of the magnitude of the primitive one. There was no exception to this rule, and it proves that the water of these particular cavities, at all events, is really due to the melting of the adjacent ice.

But how was the ice surrounding the bubbles melted? The hypothesis that the melting is due to the absorption of the solar rays by the air of the bubbles is that of M. Agassiz, which has been reproduced and subscribed to by the Messrs. Schlagintweit, and accepted generally as the true one. Let us pursue it to its consequences.

Comparing equal weights of air and water, experiment proves that to raise a given weight of water one degree in temperature, as much heat would be needed as would raise the same weight of air four degrees.

Comparing equal volumes of air and water, the water is known to be 770 times heavier than the air; consequently, for a given volume of air to raise an equal volume of water one degree in temperature, it must part with 770 × 4 = 3080 degrees.

Now the quantity of heat necessary to melt a given weight of ice would raise the same weight of water 142.6 Fahr. degrees in temperature. Hence to produce, by the melting of ice, an amount of water equal to itself in bulk, a bubble of air must yield up 3080 × 142.6, or upwards of four hundred thousand degrees Fahrenheit.

This is the amount of heat which, according to the hypothesis of M. Agassiz and the Messrs. Schlagintweit, is absorbed by the bubble of air under the eyes of the observer. That is to say, the air is capable of absorbing an amount of heat which, had it not been communicated to the surrounding ice, would raise the bubble to a temperature 160 times that of fused cast iron. Did air possess this enormous power of absorption it would not be without inconvenience for the animal and vegetable life of our planet.

The fact is, that a bubble of air at the earth’s surface is unable, in the slightest appreciable degree, to absorb the sun’s rays; for those rays before they reach the earth have been perfectly sifted by their passage through the atmosphere. I made the following experiment illustrative of this point: The rays from an electric lamp were condensed by a lens, and the concentrated beam sent through the bulb of a differential thermometer. The heat of the beam was intense; still not the slightest effect was produced upon the thermometer. In fact, all the rays that air could absorb had been absorbed before the thermometer was reached, while the rays that glass could absorb had been absorbed by the lens. The heat consequently passed through the thin glass envelope of the thermometer, and the air within it, without imparting the slightest sensible heat to either.

The liquid bubbles observed in lake ice, and those which occur in the deeper portions of glacier ice, are produced by heat which has been conducted through the substance without melting it. Regarding heat as a mode of motion, it seems natural to infer, that inasmuch as within the mass each molecule is controlled in its motion by the surrounding molecules, the liberty of liquidity must be attained by the molecules at the surface of ice before the molecules in the interior can attain this liberty. But if a cavity exist in the interior, the molecules surrounding that cavity are in a condition similar to those at the surface; and they may be liberated by an amount of motion which has been transmitted through the ice without prejudice to its solidity. The conception is helped when we call to mind the transmission of motion through a series of elastic balls, by which the last ball of the series is detached, while the others do not suffer visible separation. It may indeed be proved, by actual experiment, that the interior portion of a mass of ice can be liquefied by an amount of heat which has been conducted through the exterior portions without melting them.

Now precisely the converse of this takes place when two pieces of ice, at 32° Fahr., with moist surfaces, are brought into contact. Superficial portions are by this act transferred to the centre where a temperature of 32° is not quite sufficient to produce liquefaction. The motion of liquidity which the surfaces possessed before contact is now checked, and the pieces of ice freeze together. This appears to furnish a satisfactory explanation of all the cases of this nature which have hitherto been observed.

The particles of a crushed mass of ice at 32°, or a ball of moist snow, may, it is now well known, be squeezed into slabs or cups of ice. That moisture is necessary here, and that the same agent is necessary in the conversion of snow into glacier ice, was proved by the following experiment. A ball of ice was cooled in a bath of solid carbonic acid and ether, and thus rendered perfectly dry. Placed in a suitable mould, and subjected to hydraulic pressure, the ball was crushed; but the crushed fragments remained as white and opaque as those of crushed glass. The particles, while thus dry, could not be squeezed so as to form pellucid ice, which is so easily obtained when the compressed mass is at a temperature of 32° Fahr.