I washed by decantation with boiling water. While suspended therein, a strong solution of caustic baryta is added. The copper is thrown down as brown oxide of copper, and the double cyanides of iridium and platinum with barium are formed, which remain in solution. When the fluid reacts distinctly alkaline, after a minute or so boiling, it is filtered off, and then CO, is passed through, until a neutral reaction is obtained. Filter off the precipitate of BaCO,, and then concentrate for crystallization. The yellow platinumn salt crystallizes out first in small crystals, with a play of colors from yellow to blue, being similar in their dichroic effects to most of the other double cyanides of platinum. The iridium salt crystallizes then in larger colorless prisms, which can be easily separated mechanically from the platinum crystals. The iridium crystals I ignited in a porcelain crucible, and having thus completely broken up the compound, drew out the baryta with boiling I water. The iridium, after being well washed and ignited to completely oxidize it, was heated as before with aqua regia in a sealed tube to 200°C. The IrCl, solution obtained was again treated with absolute alcohol, and, on addition of KCl and NH,Cl, the crystallized salts were obtained.

However the compounds which were now formed were, contrary to my expectations, not the ethylen salts pure and alone, but there was a simultaneous formation of what appear to be several distinct compounds. Although in one preparation of ammonium salt analyzed, the ethylen-protochloride-salt was isolated tolerably pure, yet, in the majority of the analyses, I had to do with a mixture difficult to separate even under the lens. The readiness with which these compounds decompose when subjected to recrystallization, even although one observes the precaution of keeping the solution distinctly acid, prevents any successful purifying that way. The preparation, therefore, had to be prepared for analysis by drying between bibulous paper, and then over H,SO, or at 100°, according as I wished to determine the percentage of water or not. A certain amount of HCl from the acid solutions therefore, adhering to the crystals, made the C determinations as a rule high. Still the results approximate sufficiently, these circumstances being taken into consideration, to the formulas given, to show that several distinct compounds are here formed. My first crystallization con sisted of brownish-red octahedra, which when analyzed gave in three Cl determinations 42.18 pr. ct., 42:07 pr. ct. and 42:09 pr. ct. Cl, showing them to be simple IrCl, (KCl),, of which the percentage of Cl is 42:08 pr. ct.

I next got a crystallization of fine, sharply-formed monoclinic crystals of a reddish-brown color, which I think the analyses, although not perfectly conclusive, still go to show to be a new

compound. Several different crystallizations of it were analysed It could not always be entirely separated from the slight crust of decomposed KCl which separated out along with and among the crystals.

Preparation No 1. Large well-formed crystals, with but very little foreign matter adhering to the sides.

1567 grms. dried over H,SO, lost by heating to 100° 0170 grms. 10.85 pr. ct.:

=

gave 0525 grms. metallic iridium = 33.50 pr. ct.: gave also 2275 grms. AgCl = 0563 grms. Cl

=

35-92 pr. ct. 1610 grms. dried over H,SO, lost by heating to 100°-0172 grms. 10.68 pr. ct.:

=

gave Ir. determination accidentally spoiled:

=

gave also 2320 grms. AgCl 0574 grms. Cl = 35.65 pr. ct. This we see at once cannot be any chloride of iridium and potassium. Its luminous flame, too, when burned, shows the same. Nor do the percentages agree with the simple potassicethylen iridium protochloride IrCl2C,H,KCl+2H,O where the 48.45 pr. ct. and Cl = 26.19 pr. ct.

Ir.

=

If we reckon out the ratio of iridium to chlorine, we find it as 1: 6, showing it to be an iridium-chloride compound. The large pr. ct. of loss on heating, the small Ir. and Cl pr. cts., when compared with double iridium and potassic chloride, and the luminous flame when burned, all go to show that an organic constituent must make the additional pr. ct. If we suppose now ethylen to enter here into the union with iridium chloride, hanging itself on, we can expect from the consideration of analogous compounds that 2 atoms of the bivalent radical C, H, would join the atom PtCl..

2

2

We would have on this supposition the formula IrCl, (C,H,), (KCl), + xH2O. Now this formula with 3 atoms H2O gives Ir. = 32.93 pr. ct., Cl 35.61 pr. ct., and with 2 atoms H,0, Ir. 33.95 pr. ct., Cl = 36.71 pr. ct.; anhydrous it gives Ir. = 36.20 pr. ct. and Cl = 39·14 pr. ct.

=

=

The large pr. ct. of loss at 100° is not accounted for by 3H,0, which give only 903 pr. ct. ; but it is possible that in a compound where the organic constituent is so loosely connected with the rest, as must be the case here, a partial decomposition of the salt enters at 100° already.

In another preparation, also well crystallized, 1822 grms. dried at 100°:

gave 0678 grms. metallic iridium = 37-21 pr. ct.;

=

gave also 3063 grms. AgCl 0758 grms. Cl = 41.59 pr. ct. In another preparation, indistinctly crystallized, 0502 grms.

dried at 100°:

gave 0172 grms. metallic iridium = 34-26 pr. ct. (Probably low from HCl mechanically admixed.)

1

In another preparation, also indistinctly crystallized, 3207 grms. dried at 100°:

gave 2125 grms. AgCl

=

1268 grms. Cl = 39.53 pr. ct.

It will thus be seen the determinations, all things considered, agree close enough with the theoretical pr. cts. of the formula to make it very probable. The want of enough sufficientlywell crystallized material prevented me from making an organic combustion which might settle it definitely.

Several individual crystals of the first crystallization were very sharply and clearly formed, and I subjected them to examination under the microscope with a power of about 50 diameters. The faces were clearly to be made out. They belong to the monoclinic system, and their prevailing habitus in crystallization is a combination of the two lateral pinacoids with positive and negative pyramids accompanied by one macrodome on the ends. The faces observed in an examination of five distinct crystals were (according to Naumann) ∞ P∞. ∞ Po +P.-P. + P. - P. +P∞. -P∞.

In considering the compounds of the iridium-base with NH, Cl, we find again a mixture of crystallized salts.

Preparation No. 1 consisted of sharply crystallized needles that looked almost black, and only by transmitted light showed a brownish-green color. They were also monoclinic.

On analysis they prove, I think, to be the sought-for ethyleniridium compound.

0713 grms. dried over H,SO, lost on heating to 100° 0029

=

4.07 pr. ct.:

gave 0850 grms. AgCl

0210 grms. Cl

=

29.49 pr. ct. or

30.74 pr. ct. of salt dried at 100°.

The iridium determination was made with magnesium, and was inaccurate as before.

2

4

4

The formula IrCl,C,H,NH ̧Cl + H,O= 3675 demands Cl =28.98 pr. ct., H,O= 4.89, or anhydrous 30.47 pr. ct. Cl. Preparation No 2 was of much smaller needles and of lighter

color.

Analyses show it to be of very similar composition to the potassium salt described above.

1207 grms. dried at 100° C. :

gave 0495 grms. metallic Ir. = 41.97 pr. ct.:

2

=

gave also 2068 grms. AgCl = 0512 grms. Cl = 43.37 pr. ct. IrCl ̧(C,H,), (NÍ ̧Cl), = 502 gives Ir. 39-25 pr. ct. and Cl = 42 43 pr. ct., and supposing it to lose some of the C, H, at 100°, as stated above, IrCI, (C,H,) (NH,Cl), would give Ir. = 41.50 pr. ct. and Cl = 44.94 pr. ct.

The other two preparations of the ammonium salt analysed appeared under the lens to be mixtures of the iridium proto

chloride-ethylen salt and the iridium chloride-ethylen salt giver above.

The results were

1322 grms. dried at 100°:

gave 0609 grms. metallic iridium = 4607 pr. ct. 1715 grms. dried at 100°:

=

gave 0784 grms. metallic iridium 46.95 pr. ct. IrCl, (C,H,)(NH, CI) + H,O gives Ir. = 53.61 pr. ct. IrCl ̧ (CÍ,H), (NH,Cl), gives Ir. = 39.25 pr. ct.

2

4

The existence of the base IrCl, (C2H ̧)2,

I hope to settle definitely by renewed analyses of larger quantities.

While engaged with the preparation of the ethylen and iridium compound, the thought of the possibility of acetylen (C,H,) uniting with PtCl, or IrCl, led me to make some experiments in that direction. After a number of endeavors to form a platinum salt and analyses of the products (a detailed account of which is given in the original paper), I obtained negative results only. The existence of such a salt is highly improbable.

ART. XLV.-Directions for Constructing Lightning-Rods. From Essays on Meteorology; by Prof. JOSEPH HENRY.*

1st. The rod should consist of round iron, of not less than three-fourths of an inch in diameter. A larger size is preferable to a smaller one. (Ordinary gas pipe may be employed). Iron is preferred, because it can be readily procured, is cheap, a sufficiently good conductor, and, when of the size mentioned, cannot be melted by a discharge from the clouds. Other forms of rod, such as flat or twisted, will conduct the lightning, and in most cases answer sufficiently well. They tend, however, to give off lateral sparks from the sharp edges at the moment of the passage of the electricity through them, which might, in some cases, set fire to very combustible materials.

2d. It should be, through its whole length, in perfect metallic continuity; as many pieces should be joined together by welding as practicable, and, when other joinings are unavoidable, they should be made by screwing the parts firmly together by a coupling ferule, care being taken to make the upper connection of the latter with the rod water-tight by cement, solder, or paint.

3d. To secure it from rust, the rod should be covered with a coating of black paint.

* Smithsonian Miscellaneous Collections.

4th. It should be terminated above with a single point, the cone of which should not be too acute, and to preserve it from the weather, as well as to prevent its being melted, should be encased with platinum, formed by soldering a plate of this metal, not less than a twentieth of an inch in thickness, into the form of a hollow cone. Points of this kind can be purchased of almost any mathematical instrument maker. Usually the cone of platinum, for convenience, is first attached to a brass socket, which is secured on the top of the rod, and to this plan there is no objection. The platinum casing, however, is frequently made so thin, and the cone so slender, in order to save metal, that the point is melted by a powerful discharge.

5th. The shorter and more direct the rod is in its course to the earth the better. Acute angles, made by bending the rod, and projecting points along its course, should be avoided.

6th. It should be fastened to the house by iron eyes, and may be insulated by cylinders of glass. We do not think the latter, however, of much importance, since they soon become wet by water, and, in case of a heavy discharge, are burst asunder.

7th. The rod should be connected with the earth in the most perfect manner possible; and in cities nothing is better for this purpose than to unite it in good metallic contact with the gasmains or large water-pipes in the streets; and, indeed, such a connection is absolutely necessary, if gas or water-pipes are within the house. Electricity, by what is called induction, acts at a distance on the perpendicular gas-pipes within a house, rendering them so highly negative, the cloud being positive, as to attract the electricity from a lightning-rod imperfectly connected with the earth, or even from the air through the roof. Damage to buildings on this account is of constant occurrence. The above connection can be made by soldering to the end of the rod a strip of copper, which, after being wrapped several times around the pipe, is permanently attached to it. Where a connection with the ground cannot be formed in the way mentioned, the rod should terminate, if possible, in a well, always containing water; and, where this arrangement is not practicable, it should terminate in a plate of iron or some other metal buried in the moist ground. It should, before it descends to the earth, be bent, so as to pass off nearly at right angles to the side of the house, and be buried in a trench, surrounded with powdered charcoal.

8th. The rod should be placed, in preference, on the west side of the house, in this latitude, and especially on the chimney from which a current of heated air ascends during the summer

season.

9th. In case of a small house, a single rod may suffice, provided its point be sufficiently high above the roof; the rule