ing to sunlight. The results obtained in this way are given in figs. 1-7.

FIG. 3. Framework after flow.

= 1.59cm. The faces in

FIG. 3. The cube was compressed through in. contact with the plates extended to 2 in. = 5.24cm on a side.

FIG. 4. Framework before flow. FIG. 5. Same framework after flow. FIGS. 4, 5. The cube was compressed through in. = 2.38cm. The faces next to the plates extended to 28 in. = 6·03cm on a side. In Fig. 5, one cross-wire is missing next to the right hand end.

The

FIGS. 6, 7. The cube was compressed through 1 in. = 3.26cm faces next to the plates extended to 2 in. = 6.51cm on a side. In Fig. 7, the cross-wire at the extreme right hand end became detached during the treatment with H2SO4, and is missing.

It will be observed that the wires became considerably flattened in the last stages of the flow, and that slight wavy irregularities in their form appeared, probably on account of imperfect homogeneity of the material. It is possible, however, to follow in the figures the general character of the motion, from the first stages of the flow to the last. At first the flow does not differ in any essential respect from what might be expected in view of the shape taken by the surface of the cube. But as the flow proceeds, the wires near the middle of the cube take the form of curves with two points of inflection, as shown in fig. 5, and less distinctly in fig. 7. In figs. 7, 5, and 3, a shearing of the cube parallel to the plates of the testing machine is apparent; the upper part has moved to one side, and the lower part to the other. These results are not surprising, because compression without such shearing would represent an unstable state of motion; if any cause should slightly displace the lower part of the cube with respect to the upper part, perpendicularly to the direction of the compression, then further compression would tend to increase such a displacement.

In view of this complication, it does not seem worth while to attempt to apply to the compression of a cube the mathematical theory of the flow of solids, as developed by Tresca and Saint Venant.*

University of Chicago,
September, 1901.

* Comptes rendus, lxvi, pp. 1027-1032, 1244-1246, 1305-1324; lxviii, pp. 221-237, 290-301; 1xx, pp. 309-311, 473–480; etc.

ART. XIX. On the Occurrence of Monazite in Iron Ore and in Graphite; by ORVILLE A. DERBY.

A SMALL specimen of magnetic iron ore presented by Mr. John Gordon of Rio de Janeiro, from the fazenda Catita, on the lower Rio Doce in the state of Espirito Santo, presents a number of interesting features, among which is the occurrence of numerous and comparatively large grains of monazite in the mass of the ore. The ore fragment consists of a coarsely crystalline mixture of magnetite and ilmenite with adherent remnants of kaolinized feldspar and biotite, which show it to have been a segregated mass of oxides in the midst of a coarsely granular rock, probably a mica-syenite. The powdered ore, freed from the iron oxides by the horseshoe and electro-magnet, gives an extremely abundant residue of rather coarse fragments and well crystallized grains of corundum, monazite and zircon, and in microscopic slides these grains are found to the number of a dozen or more in the area of an ordinary preparation. They occur isolated in the mass of the oxides, but are more abundant in and about flakes of biotite when these are present. Of the three minerals, monazite is the most abundant and the most generally distributed, appearing in both the magnetite and ilmenite. Other interesting accessories that are confined to the magnetite, where they appear as delicate net-like partings in the twinning planes (something like the plates of tænite in meteoric irons), are a green spinel and a translucent brown titanium mineral. The ilmenite also gives on etching irregular bands showing it to be composed of a mixture of two substances of different color, and degree of solubility in hydrochloric acid. This and other interesting features of titaniferous iron ores from this and other localities will be more fully discussed by Dr. Hussak.

A specimen of graphite has recently come to hand from the region of the river Jequitinhonha in the state of Minas Geraes, which gives on washing a very abundant residue of heavy yellowish fragments, rarely crystals of recognizable form, that on microscopic and chemical examination prove to be monazite and zircon, the former greatly predominating. The only other recognizable element of the residue is a dirty white opaque titanium mineral that seems to be a pseudomorph after mica. The compact graphite is transversed by thin stringers of a decomposed micaceous mineral which also occurs in small isolated rounded patches, but these afford no more, if as much, monazite as the purer portions of the specimen. Several isolated flakes of graphite with an included grain of monazite were obtained. From the rarity of perfectly formed crystals

or of the rounded grains in which the mineral usually appears, the monazite seems to be in a state of strain in virtue of which it goes to pieces in the process of crushing and washing. On testing in a borax bead the oxalates precipitated from a solution of the residue, Dr. Florence obtained beautiful crystallizations of both cerium and lanthanum, the latter appearing much more abundantly and readily than in the many other samples of monazite that he has examined in this way. From this circumstance it may be concluded that the mineral presents some peculiarities of composition, but material is not at hand for a verification of this point.

The locality from which the specimen comes was visited in 1880 by Dr. Costa Sena, the present director of the School of Mines of Ouro Preto, who reports the occurrence of loose masses up to 100 kilograms in weight and of a vein from half a meter to a meter in width in decomposed granitoid gneiss in the bed of the Corrego do Emparedado, affluent of the small river São Pedro, which enters the Jequitinhonha from the left some 60 to 70 miles below the town of Calhau (Arassuahy). An analysis made at that time gave 85% of carbon, 47% of volatile matter and 7·2% of ash. Judging from the present specimen, which probably was about the same composition, the ash is composed for the most part of phosphates of the cerium group in the form of monazite.

Another specimen of graphite of similar appearance and mode of occurrence, from near São Fidelis in the state of Rio de Janeiro, presents the same phenomenon of an abundance of monazite as the almost exclusive non-carbonaceous accessory. On the other hand, several specimens of graphitic schist that have been examined give an abundant residue of titanium minerals (rutile in some cases, ilmenite in others), but no minerals of the rarer elements. This circumstance and the occurrence in the schistose types of graphite of a great amount of sericitic mica indicates a difference in the mode of origin of the two types of graphitic rock corresponding to the differences in their mode of geological occurrence. Unfortunately no specimens are at hand for verifying to what extent monazite is a characteristic accessory of the graphite occurring in gneiss and granite. As the two specimens examined, taken by chance, have shown it in relative abundance, it may be suspected that it will be found to be rather generally distributed. If so, its significance from a genetic point of view hardly needs to be mentioned.

São Paulo, Brazil, December 24, 1901.

ART. XX.-The Molecular Weights of some Carbon Compounds in Concentrated Solutions with Carbon Compounds as Solvents; by CLARENCE L. SPEYERS.

IN some preceding papers, it has been shown that the equation

expresses the vapor pressures of mixtures of liquids miscible in all proportions better than does the equation

The latter gives absurd values towards the limits of concentration the former gives reasonable values throughout the range of experiment. The failure of (2) is hardly to be attributed to experimental error,t because in that case the molecular weights should show far more irregularity than they do.

Moreover with such simple assumptions regarding molecular weights as are continually being made for non-volatile solutes, it was possible to plot boiling point curves for mixtures of two liquids soluble in each other in all proportions and to state in a general way when a mixture might have a maximum boiling point.+

Last year, J. von Zawidski§ applied Margule's equation

where a denotes the fraction of a gram-molecule of one liquid and 1-x denotes the fraction of a gram-molecule of the other liquid, to a mixture of two liquids miscible in all proportions. In this differential form, the equation seemed to be unsatisfactory, and to integrate it constants of uncertain value are introduced complicating the theoretical investigation. Moreover, the equation will not agree with experiment unless molecular association is granted. Taken altogether, Margule's equation is more complicated than equation 1, far more so, and in no case does it give results more concordant with fact than equation (1) does. For one mixture, that of acetone and chloroform, equation 1 fails. But so also does Margule's equa

Journ. Phys. Chem., ii, 347, 362, 1898; Journ. Am. Chem. Soc., xxi, 282, 1899.

Bancroft, Journ. Phys. Chem., iv, 224, 1900.

This Journal, ix, 341, 1900.

$Zeitsch. Phys. Chem., xxxv, 129, 1900.

AM. JOUR. SCI.-FOURTH SERIES, VOL. XIII, No. 75.-MARCH, 1902.