of action will be 1, 3, 9, 27, etc. They further state that this law of action holds good up to about three per cent., and that for solutions weaker than this, they have demonstrated it through a range of from 1 to 1,000,000.

They also give for the law a mathematical expression which seems to have been defectively printed, for when the typographical errors are corrected I find it to be expressed by

where C is the rate of chemical action, p is the proportionate quantity of salt, and K is a constant. In the particular example given in their article they employed a plate of copper suspended in a solution of argentic nitrate. But the authors do not give any explanation of the law which they have discovered.

The following question at once arises, Is the rate of metallic precipitation for varying strengths of the solution due to the action of chemism alone, or, is the operation of chemical affinity compounded with one or more purely physical forces? and if so, what is the relative part contributed by each?

Messrs.Gladstone and Tribe point out that there are two currents, or mass movements produced. One, a light blue ascending stream of a dilute solution of silver and copper nitrates; the other, a deep blue descending current of copper nitrate containing about three times as much NO, as the main solution. Obviously the first step in an investigation of their results should be to bring the fluid against the plate with a uniform velocity. To accomplish this a piece of copper was slightly curved and fastened to an arm which carried it around in a horizontal circle of about one inch radius with a uniform speed of five and a half revolutions per second, which gave therefore a linear velocity of the metal through the liquid of nearly three feet per second. The strip was suspended from the arm so as to hang vertically, and the area of its surface was known. The results are given in the appended table in the columns marked 2 and 3. It was at once apparent that the absolute rate of action was much greater than when the metal was stationary, as in Messrs. Gladstone and Tribe's method where the only mass motion of the liquid relatively to the plate was caused by gravitation. An inspection of columns 2 and 6, where their method was followed, will show this.

In column 3 are given the quotients of the actual weights divided by the percentages of silver nitrate. It is apparent that these

numbers tend toward a constant, and, in this respect, they show a marked contrast to the numbers in column 7.

The experiments were next varied by varnishing the copper on its convex side and holding it rigidly against the side of a beaker while its bare concave face was constantly swept by a rotating cylindrical brush, driven at uniform speed by clockwork. The results are given in columns 4 and 5, where, to secure a greater range for comparison, a different order of per cents was employed, and it is evident from column 5, that the absolute rate is less than for the rotation experiments, but is still much greater than for the gravitative record; also it is apparent that the quotients here also tend to approach a constant [.6].

The first two numbers in columns 3 and 5 depart widely from the others. This is due, I believe, to the necessity of removing the silver each time the copper was weighed to determine how much of it had been dissolved. When the solution is very weak, the silver crystals are so minute that it is practically impossible to remove them completely, while for larger crystals, made by stronger solutions, there is no such difficulty.

Now when the extreme difficulties of securing absolutely uniform rotation, of detaching the copper at a given instant, and of preventing accidental cross currents are considered, these numbers show (since they are the means of many trials) that the chemical action varies directly with the quantity of salt in solution, when the supply of fresh liquid to the surface of the plate is constant and independent of its strength. Within the limits of these experiments, the following law would appear to be true. In metallic precipitations of dilute solutions, the chemical action varies directly as the mass, i.e., percentage, of the dissolved salt.

The experimental work of Messrs. Gladstone and Tribe is above criticism. I made many determinations by their method, with the sole result of confirming their entire accuracy. How then can their law be reconciled with the above statement?

The explanation I venture to offer is the following:

When the plate is suspended in the liquid, a film of copper nitrate is at once formed against it. Now the access of fresh silver nitrate depends on the removal of copper nitrate; this film being heavier than the main solution begins to fall (a part also rises) and thus drags in fresh liquid from above to attack the copper. The rate at which it falls depends on its density, that is

on the quantity of copper dissolved, but it is resisted by the inertia of the surrounding fluid which it must displace.

Now in a second experiment let the percentage of silver nitrate be doubled; then, by the statement made above, the quantity of copper dissolved in a very small portion of time will also be doubled, and its density will be greater. The increase of density may be found from the tables of specific gravities of copper nitrate solutions. Storer's dictionary of solubilities gives the following:

We see that the increment of the density rises a little more rapidly than the increase of the salt, so that the density of a four per cent solution, for example, is 2.11 times that of a two per cent solution. Therefore the film of copper nitrate will be more than twice as dense when the percentage of silver nitrate is doubled. The rate of motion of the film will, however, not be twice as great; for bodies, moving with slow velocity through a liquid, are found. to experience resistances which vary as the square of the velocity. The density of the film has become more than twice as great, consequently the attraction between it and the earth has more than doubled. It is also evident that though the specific gravity of the silver salt is increased, still, the increment of the density of the copper solution over that of the silver solution will be more than doubled. Hence its velocity will become not two, but as the square root of two, which is 1.41 +: but as the relative density (or the relative increment) is really more than two, its square root will approach near 1.5.

The observed rate of action will be composed therefore of the two ratios, viz., twice the chemical action because of chemism, multiplied by one and a half because of gravitative action; or, when the per cent of silver salt is doubled, the quantity of copper dissolved will be tripled. But this is Gladstone and Tribe's law.

The above hypothesis leads to this conclusion. If the "2-3 law" is composed of two factors, one of which is chemism, and the other gravitation; then if another accelerating force could be substituted

for gravitation the law ought to change its form, if that force could be made a large one, for, the rate of movement of a body through water is determined by resistances which increase much faster than the square of the velocity when the latter is no longer small. Accordingly I mounted the precipitating vessel on an arm which permitted the whole to revolve in a vertical plane and with a velocity which gave a centrifugal force thirty times greater than the value of the acceleration of gravity. With this large force it was certain that the currents ought to increase much more slowly than in the ratio of the square root of 2. An inspection of columns 8 and 9 will show the results. In column 9 the numbers do not tend to approach a constant but slowly increase. They resemble in this respect the gravitation numbers in column 7, only their rate of increase is slower as it should be by theory. A mathematical analysis of these numbers, taking the 30 per cent one as a basis, agrees very closely with those calculated by the formula— that doubling the per cent of silver increases the rate of precipitation to 2.25 instead of to 3.

In conclusion, I think it is probable that the true law of chemical action, where one metal precipitates another, should be that the time during which one atom replaces another in a compound molecule is fixed, and hence that the rate of total chemical action varies directly as the mass of the reacting body in solution.

PRELIMINARY ANALYSIS OF THE BARK OF FOUQUIERIA SPLENDENS. By HELEN C. DE S. ABBOTT, Philadelphia, Pa.

In the published proceedings of the Mexican Boundary Survey of 1859, conducted by General William H. Emory, are found numerous references to Fouquieria splendens. No region of equal extent presents more marked illustrations of the relations of the vegetation of a country to its topography and geology than that lying along the Mexican boundary line. The traveller traversing the desert table-lands will not fail to unite in his recollections of these tracts the dull foliage of the creosote bush, the palm-like Yucca, and the long thorny wands of the Fouquieria splendens. The vegetation of the El Paso basin and the Upper Rio Grande

valley is described as strikingly different from that of the immediately adjoining country: new and strange plants are seen on every side. Upon the table-lands many plants grow not to be found in the more fertile valleys; among these is Fouquieria, a tree locally known by its Mexican name ocotilla. A full description of the appearance of the plant is given in the Mexican Boundary Survey; also one in an article by Edward Lee Green. The latter author describes Fouquieria splendens in these terms: "It is a splendid oddity and not more odd than beautiful, flourishing in great abundance in many places. It grows to the height of from eight to twelve feet, and in outline is quite precisely fan-shaped. The proper trunk, usually ten to twelve inches in diameter, is not more than a foot and a half high. A few inches above the surface of the sands this trunk abruptly separates into a dozen or more distinct and almost branchless stems. These simple stems rising to the height of eight or ten feet gradually diverge from one another, giving to the whole shrub the outline of a spread fan. Each separate stem is clothed throughout with short gray thorns and small dark green leaves, and terminates in a spike, a foot long, of bright scarlet trumpet-shaped flowers. The stems are not so thickly armed with thorns but that they can be handled if grasped circumspectly, and being very hard and durable, as well as of a convenient size, they are much employed for fencing purposes about the stage stations and upon the ranches adjoining the desert." The author states: "Give a skilful Mexican ocotilla poles and plenty of raw hide thongs and he requires neither nail nor hammer to construct a line of fence, which for combined strength, neatness and durability fairly rivals the best work of that kind done in our land of sawmills and nail factories."

The plant is botanically described under order Tamariscineæ, tribe III, Fouquiereæ, new genus and species. For other sources of information see A Tour in New Mexico ;3 and in Planta Wrightianæ Texano-Mexicanæ. The writer has not been able to find any notice of chemical studies made upon it.

The specimens of ocotilla, at the writer's request, were collected and transmitted from Lake valley, Southwest New Mexico, through

63.

1 Botanizing on the Colorado Desert, American Naturalist, 1880. "Bentham and Hooker. Genera Plantarum.

By Dr. N. Wislizenus.

4 Gray, Smithsonian Contributions to Knowledge. Vol. iii. Part 1, p. 85 and Pt. ii, p.