According to Desmond Fitz Gerald, M. Am. Soc. C. E., in a paper recently read at the annual convention of the society, the observation of the temperature of the water in lakes and reservoirs is attended with more difficulty than is generally supposed. Hitherto the taking of readings at any considerable depth has been rendered difficult and tedious on account of the unsuitability of the ordinary mercurial thermometer for such work. The invention of the thermaphone by Messrs. H. E. Warren and O. C. Whipple enables the observer to take in a few minutes a more accurate reading than was formerly possible after an hour s careful work. The thermaphone is based upon the principle of the Wheatstone bridge, and it enables the temperature to be read at the surface of the water, the two metals which form the circuit being suspended at the desired depth. The two arms which complete the circuit at the surface are connected with a telephone which takes the place of the customary galvanometer. The theory of this very sensitive and accurate instrument is based upon the fact that different metals have different electrical temperature coefficients. A circular slide wire is connected to the two coils of dissimilar metal, which are lowered to the desired depth by means of two leading wires. This slide wire is wound around the edge of a disk, which carries a dial, graduated in degrees of temperature. A leading wire connected with the junction of the two metal coils connects with a telephone receiver and terminates in a radial contact arm, which travels upon the above mentioned disk. The ends of the slide wire are put in circuit with a battery. In reading the temperature, the radial contact arm is moved back and forth over the dial, and the telephone is held to the ear. The buzzing sound in the telephone increases or decreases as the hand passes a certain point on the dial. By continually moving the hand, a point will be found at which the sound ceases altogether. The reading at this point indicates the temperature of the distant coil. This instrument is so accurate that its results can be depended upon to much less than 0 1 F., and a series of temperatures throughout the vertical can be taken with an allowance of about a minute for each point observed. Surface Temperatures.—During the winter, from the latter part of December to the breaking up of the ice in the spring, the temperature of the water under the ice is 32 F. The water then warms at a uniform rate to 72 F. in the middle of June. From that time to the middle of August it varies between 73 and 78, and then falls regularly to 37 in the middle of December. Bottom Temperature.—In a pond less than 25 feet deep the bottom temperature varies very little from that at the surface. In the deeper lakes very interesting phenomena occur, which have an important bearing upon the question of domestic water supply. The observations were taken in connection with the Boston water works at Lake Cochituate. Th e point of maximum densi ty of fresh wat er iS 39 2 F. This is about the temperature of the bottom of the lake when the surface freezes. The several strata lie in their order of density, decreasing gradually until within a few feet of the surface, when they suddenly fall to the freezing point adjoining the ice. T he body of water remains unchanged throughout the winter. At the breaking up of the ice, the surface water warms up to the temperature of the bottom layers ; the whole body is thrown into unstable equilibrium, and circulation takes place from top to bottom. As soon as the surface is 5 F. warmer than the bottom, circulation ceases. Although the temperature of the surface continues to rise, the bottom remains at exactly the same temperature throughout the long period of stagnation, covering about seven months, during which time it varies only a few tenths of a degree. From this it is evident (1) that the agitation set up by the winds at the surface does not penetrate very deep (experience shows fifteen feet to be about the limit); (2) that there are no convectional currents at work to effect a change of temperature ; and (3) that water is such a poor conductor of heat that the hottest sun s rays are not perceptible at the depth of sixty-five feet. Weekly observations of temperature in Lake Cochituate for a period of four years show that the surface agitation by the wind keeps the water at an even temperature for the first ten feet of depth, and that below fifteen feet the effect is very slight. The Effects of Stagnation.—The deeper, quiescent layers of water gather the organic matter from the waters above, and decay goes on until the oxygen is used up. The water becomes dark in color and acquires a disagreeable smell. Commenting upon these facts, Mr. F. P. Stearns stated that these lower strata of water, which are unable to get any fresh supply of oxygen from the air, accumulate free ammonia and other solid and gaseous products of decomposition. Hence it is desirable that the domestic supply should be taken from near the surface and waste water drawn off from the bottom. In this way the evil effects of summer stagnation may be partly overcome and the whole body of water improved at the autumnal overturning. It is possible in the summer to sink a bottle to the 1 bottom of Lake Cochituate and bring up ice-cold water, and, at the same time, fill another bottle with water from the surface that is 80 in temperature. In view of this great difference in temperature and the purity of the surface water, the author of the paper suggests that any one living near a deep lake could obtain very pure ice-cold water during the summer months by taking it from the surface and leading it through a coil of pipe placed in the cold stratum of water at the bottom. Observations of Lake Superior, taken in August, show a bottom temperature of 38 8 F. in 158 fathoms, the surface temperature being 50 to 53 P. Prof. Le Conte. in August, 1873, found Lake Tahoe, in Cali ornia, to be 39 2 at 1,506 feet; 41 at 772 feet, and 67 at the surface. Nine soundings, taken in Lake Thun, in 1848, to a depth of 550 Swiss feet, show a mean temperature of 40-7 . The Lake of Geneva, which is 1,000 feet deep, shows a mean temperature of 41 2, as the result of seven years of observation. As the result of his own and other observations, the author arrives at the conclusion that in a lake of the first order, like that of Geneva, the winds produce a mechanical mixture of the layers to a considerable depth below the surface; the smaller the lake the less these mechanical effectsare felt, but that this heating is not due to conduction seems to be proved by the fact that, at 65 feet depth, conduction has no effect in seven months time on the bottom temperature of Lake Cochituate.