the rest of the substance is ready to pass. Thus in the cooling of water below the freezing point, a bit of ice causes the immediate congelation of the entire mass; in the superheating of water, the presence of the smallest globule of saturated vapor below the surface of the liquid, brings about explosive boiling. Knowing this, it is easy to surmise the nature of this difficulty of beginning the change of state; the change of state is a process of fluid diffusion involving the presence of the substance in both states, and the reluctance to begin the change occurs whenever the substance is present in one state only. In the super-saturation of salt-solutions this same difficulty of beginning the change of state occurs; and the difficulty disappears when a particle of the substance in the solid state is introduced into the solution; just as in the other cases it vanished so soon as the substance was present in the two states. This indicates that the process is the same in the two cases; in a word, that, as we had concluded from other considerations, solution is a process of fluid diffusion.

The other phenomena connected with the change of state admit of equally simple explanations, if we adopt one more hypothesis concerning the nature of liquids. I feel the more confidence in making use of this hypothesis, since Professor Clausius has suggested and defended it in a recent paper in Poggendorff's Annalen. According to this view, the liquid molecule (meaning again by molecule that portion of the liquid which moves independently) is an aggregation of the gas-molecules of the substance; and its mass is a function of the temperature-decreasing as the liquid is heated. Andrews' critical point is then for each liquid, the temperature at which the liquid and vapor molecule are identical. At lower temperatures, the ability of liquid and vapor to exist, side by side, exhibiting such different properties, is explained by the difference in the size of their molecules. Under this hypothesis, the total energy necessary to the vaporization of a liquid is divided into two portions: that involved in raising the mean velocity of the molecules, and that consumed in breaking up the larger molecules into smaller ones. This latter portion is in common parlance, the latent heat of vaporization; it decreases with rise of temperature, and at the critical point becomes equal to

zero.

Under this hypothesis, finally, we can account for the high rate of evaporation of gases in solution and the low rate of salts in so

lution. During evaporation it is chiefly the molecules of highest velocity which escape - hence the lowering of the temperature of the liquid from the evaporation. If the liquid molecules of a gas solution are of much greater mass and of much smaller mean velocity than the gas molecules, the absorption of the latter will raise the mean velocity of the mixture, and cause the well-known rise of temperature due to absorption, but the small gas particles will retain their own high velocity instead of imparting them to the large liquid molecules.

Were gas molecules and liquid molecules of nearly equal mass, a distribution of velocities would occur, since by a single collision the entire velocity of a gas molecule could be imparted to some liquid molecule; but it is evident that since no gas molecule impinging upon a very much larger liquid molecule can sensibly change the velocity of the latter, and since in the long run the sum of the velocities imparted to such a liquid molecule by such collisions is equal to zero; there can no such distribution of velocities take place. The gas molecules in the mixture will therefore retain their high velocity and will evaporate more rapidly than the liquid molecules. In the same way, if the molecules of a salt in solution are vastly larger than the liquid molecules the liquid molecules will maintain their higher velocities and the liquid component of the solution will evaporate most rapidly.

This fact, that in salt and gas solutions the different components evaporate separately and independently and at very different rates, implies their existence in the solution in independently moving molecules. The hypothesis of a complex molecule offers no such simple explanation of the phenomena connected with solution; therefore is not tenable. From the other hypothesis the classification of the process of solution together with the other processes by which matter changes state, under the head of fluid diffusion, follows almost as a matter of course.

NOTES ON WATER ANALYSIS. By A. A. BRENEMAN, of Ithaca, N. Y.

1. Significance of the chlorine test.-Notwithstanding all that has been said upon the subject, it seems to me worthy of remark, that the determination of chlorine in well waters is of the greatest importance, under certain precautions, and in many cases should alone condemn the water. My own experience in more than two hundred determinations of chlorine, in the waters of the gravel district of Central New York, determinations made together with the ordinary analysis by Wanklyn's method, sustains this view.

2. Change of Organic Matter by Aeration and Light.-A sample of peaty water, deep brownish in color, and yielding 0.44 per million of albuminoid ammonia, gave after flowing one and a half miles, with slight fall, .25, and at the end of three miles .11, the color passing from deep brown to almost nothing, at the same time.

The aeration of suspended matter in streams should in general lead to a more liberal interpretation of its influence, than in the case of the same kind of matter in well waters, which are deprived of aeration. Too little importance is attached to influence of animal life and strong light as purifying agents, in running streams.

3. On the Volume of Water to be used in Water Analysis.—The use of 100 cc. for each distillation by Wanklyn's method is preferable, because of the greater convenience and quickness of the operation, and also because of increased delicacy in Nesslerizing. Increased delicacy results from the use of smaller tubes, which give a length of column greater in proportion to volume of distillate used in Nesslerizing.

4. An Improved Nessler pipette-stirrer.- The pipette used for the Nessler test, as here shown, is distinguished from the ordinary bulb-stirrer by the very short tip below the cord or bead, which latter fits closely to the tube in which the test is made. The wide opening prevents spirting of liquid from the tube as the stirrer is raised and lowered rapidly. The short tip permits agitation of the lowest portion of liquid in the tube.

5. On the use of Common Yellow Glass as a Color Standard in Nesslerizing.-The advantage of colored glass is its permanence and convenience of application. The glass in ordinary use is

suited only to the determinatlon of high colors. Thin glass, of the same composition, might be made to satisfy all necessities of the case. The standard slip of glass is marked with a diamond, and the simultaneous use of a bundle of slips, of different values, permits the accumulation of any desired degree of intensity in color.

ON COLORS FOR SALT-GLAZED POTTERY. By A. A. BRENEMAN, of Ithaca, N. Y.

IN continuation of some experiments, reported at the last meeting of this Association, I have succeeded in obtaining in the ordinary stone-ware kiln, all of the colors shown upon the experimental tile here shown with many other tints and shades, not heretofore produced, to my knowledge, on this kind of ware. I have little doubt that a complete palette of colors may be obtained for the decorator in this field. The number of experiments made is so great, and the results are so various, that I shall not attempt a description of them at this time, confining myself for the present to a statement of a theory of the action of the saltglaze in its relation to mineral colors. This I believe, to depend upon a circulation by convection currents in the thin layer of sodium silicate resting upon the ware, after salt has been thrown into the kiln. As chlorine can have little affinity for the silicious matter of the glaze, while it is capable of volatilizing many of the metals which, in one form or another, are used as pigments, it would seem that the destruction of color, well known in the case of ordinary pigments, may be prevented by giving such consistency to the color-holding medium, that convection currents shall be impossible, at the temperature of the kiln; while at the same time this medium shall have mobility sufficient to permit of that migration of molecules, necessary to a uniform distribution of the color. Upon this theory my results have been obtained.

ON THE DEFICIENCIES OF METEOROLOGICAL WORK IN DATA OF VALUE TO AGRICULTURE, AND MEANS FOR SUPPLYING THEM. By Wм. MCMURTRIE, of Washington, D. C.

WHAT I mean by deficiencies of meteorological work in this connection is the want of data in the published records, concerning many of those conditions recognized by agriculturists and vegetable physiologists, to have an important influence upon the normal development of animal and vegetable life; and that the observations of the present time are not such as will supply all the data required in the study of meteorology, in its relations to agriculture, and especially to the condition of the crops. The want of such data, I felt most keenly in my work of last year, a portion of the results of which I had the honor to present to the association, at the Saratoga meeting. In that work, the object was to determine the limits to the conditions to which the crop in question might be subject, and reach such development as would give a profitable return to producer and consumer; and while the results obtained were interesting to me at least, it was difficult to make them so thoroughly marked as would have been possible, if more complete data, such as I will mention later on, had been obtainable.

The records published give detailed data concerning barometric pressures, temperatures observed in the shade, rainfall, relative humidity, direction and force of the wind and conditions of the sky, expressed in percentages of cloudiness, and these are valuable in their way, and for the end to which the observations were taken but they by no means cover all that must be recognized in scientific and practical agriculture. Radiated as well as diffused heat, diurnal and monthly distribution of rainfall, evaporation of moisture, prevalence of dews, fogs and frosts, and tension of atmospheric electricity are also important and necessary. Light, that most patent factor in vegetable development, has been almost wholly neglected, except in the record of the condition of the sky, as expressed in percentage of cloudiness, which, in this connec tion, is as much a matter of accident as design.

Practical agriculturists have almost completely ignored the science of meteorology in the prosecution of their labors, repeatedly making attempts at the introduction of new crops into sections wholly unsuited to them on account of climate; and on the other hand, in the work of meteorologists, although agriculture