the apparatus was filled with a mixture of the respective solutions and distilled water in the proportion of five cc. of the solutions to one liter of water. Trials with fifteen cc., and even ten cc. of the solutions to one liter of water, failed on account of over-nutrition, the seedlings refusing to grow and finally dying.

The plants employed were dwarf peas. The seeds were made to germinate in sterilized sand, moistened with distilled water. When the seedlings were four or five inches in height, they were taken from the sand, the roots washed and the plants transferred to the culture jars.

Three sets of apparatus, as shown in the illustration, were provided. The first set contained the complete nutrient solution. The second set contained solution No. 2, as described above, without nitrogen.

The third set contained the same solution, but in addition the seedlings before being placed into the jars, were inoculated with tubercle germs, by immersing the roots for a few minutes in a cold water infusion of a soil, in which peas had been grown.

As might be expected, the plants in set No. 1, containing the complete nutrient solution, were healthy and vigorous, and produced flowers and fruit.

The plants in set No. 2 grew rapidly for a time, but were weak and sickly in appearance. In about ten days the leaves began to turn yellow and to show the effect of nitrogen starvation. A few small blossoms finally appeared, but this caused the speedy death of the plants. The roots developed very rapidly, coinpletely filling the jars. They were doubtless in search of nitrogenous food.

The plants in set No. 3 grew normally for about ten days, when they began to show the effect of nitrogen starvation in a marked degree. No tubercles could be observed on the roots. About the thirteenth day the plants began to recuperate, the leaves assumed a normal green color, and from this time on the growth was vigorous and normal. On the fifteenth day the tubercles were first observed, but they had then attained considerable size.

The plants, like those of set No. 1, produced flowers and fruit. Investigations in this direction will be continued.

[CONTRIBUTION FROM THE JOHN HARRISON LABORATory of ChemISTRY,

No. 25.]

DERIVATIVES OF SILICON TETRACHLORIDE.'

BY JOSEPH F. X. HAROLD.
Received December 1, 1897.

GENERAL REMARKS UPON COMPOUNDS OF THE HIGHER HALOIDS OF THE ELEMENTS OF GROUP IV.

THE

HE review of the reactions of these tetrachlorides furnishes some interesting data from which inferences of importance may be drawn. There are, however, numerous problems still unsolved, and many reactions that yet demand study, before any broad comparison of the behaviors of these elements can be definitely made, or any conclusion as to the influence of the atomic weight or metallic character on the reactivity of their tetrachlorides established. No general rule could be deduced from the data already collected, which would enable us by a logically drawn analogy, to predict their behaviors in certain. uninvestigated reactions.

Carbon and silicon, the first two elements of the group, are by their position more or less isolated, and removed from too close a comparison with any of the sub-group elements, and their reactions are best studied with a view to determining or increasing their own already well developed similarities, without seeking to broaden to too great an extent their relation to other elements of the group. The chlorides of silicon and carbon may be said, however, in a general way, to possess in common with the other group members, a distinctly acid nature, and to exhibit the power of combining with bases to form stable, well defined compounds. This behavior may be predicted for even the unstudied chlorides of germanium, lead and thorium, since every element thus far investigated has in its tetrachloride form thus deported itself. Ammonia, the substituted ammonia derivatives, such as methylamine, toluidine, urea and amido bodies of a distinctly basic character, may here be said to react similarly with all. With acid amides and bodies of a distinctly acid nature, no such general rule may be adopted, and a series of reactions occur which seem to be conditioned by no rise or fall in atomic weight

1 Thesis presented to the University of Pennsylvania for the degree of Doctor of Philosophy.

or by position and group relation of the elements. Titanium, one of the elements of the sub-group having a less metallic character, and tin of the opposite sub-group and of a distinctly metallic nature and a high atomic weight, represent the most reactive of the elements of Group IV, in their tetrachloride form. Other than this, there seems to be little relationship between them, being distinctly separated by such considerations as place them in opposite groups. Titanium, the element of lowest atomic weight of the first sub-group, is in its quadrivalent form most reactive. Its tetrahaloid forms derivatives with hydrocyanic acid, cyanogen chloride, and the oxides of nitrogen. It thus has the power of holding acid as well as basic compounds in combination. Zirconium, on the other hand, the element next in succession to titanium, exhibits no such reactions, and its combinations seem limited to compounds of a basic character, as work now in progress in this laboratory indicates. We might then assume the rule "that in the first sub-group the reactivity decreases with a rise in atomic weight, being highest in titanium," with more or less reason. In the case of the opposite sub-group, however, the paucity of data forbids the formulation of any such law. Tin we know to be exceedingly reactive in its quadrivalent haloids, but germanium, the element which precedes it in the sub-group, whose reactions, if studied, would make some such general statement possible, has been but little investigated with such an end in view.

The subsequent chlorides of thorium and lead are being studied at present, and the determination of their reactions will show the relevance of their weight relationship to their behaviors, and make possible a more comprehensive and thorough comparison of the actions of the tetrachlorides of the fourth periodic group. The accompanying table represents the behavior of the several tetrachlorides towards various reagents. The reactivity is indicated by the letter "r," and the failure to form compounds with these reagents by the character "o."

The work was entered upon chiefly with a view to developing some new analogies between the behavior of silicon tetrachloride and that of the tetrachlorides of the other members of Group IV, or any characteristic differences which might prove noteworthy. The hope of the investigation was, that by some well defined series of reactions observed therein, a new means might be found of isolating the members of the group, in their higher forms of combination, from one another or from members of other groups. The investigation is all the more demanded, since but little work has been done in this field, the action of nitriles having been studied only with the tetrachlorides of titanium and tin.' The action, therefore, of nitriles on silicon tetrachloride was a needed expansion of the subject, and this, together with the actions of the chlorides of sulphur and phosphorus, cyanogen and chloride of cyanogen, nitrogen dioxide and nitroxyl chloride, upon silicon tetrachloride, constitute the experimental part of the present research, while the demonstration of the inactivity of the latter, as compared with the behaviors of tin and titanium tetrachlorides, determines its scope.

The second part of this article deals with the reactions between silicon tetrachloride and amines of the benzene series. The behavior of these bodies towards higher haloids of the fourth group has been investigated only in the silicon derivatives, and

1 Henke: Ann. Chem. (Liebig), 106, 281; Shinn: Thesis, University of Pennsylvania, 1895.

here the reaction cannot be said to be determined, since the investigators in this field are at odds in their results, and it is hoped that this work will aid in fixing the exact constitution of the products of the reaction of silicon tetrachloride and aromatic amines.

The silicon tetrachloride was made by the action of dry chlorine on silicon, at a low heat, the resulting vapors being collected in chilled condensers. Any free chlorine was removed from the product by shaking with mercury, and the silicon tetrachloride further purified by fractional distillation, till it exhibited the constant boiling-point, 57°-59° The silicon used

in this preparation was made by the reduction of very fine sand, free from iron, by magnesium powder, according to the method recommended by Gattermann.

THE ACTION OF FORMONITRILE ON SILICON TETRACHLORIDe.

Dry hydrocyanic acid gas conducted into cooled silicon tetrachloride produces no change; no product is formed and the silicon tetrachloride boils at the usual temperature. Wöhler' had prepared the compound TiCl,.2HCN, while Klein got a corresponding tin derivative, as did also Shinn. the attempt to combine formonitrile and with the result above indicated.

ACETONITRILE.

These results led to silicon tetrachloride

Henke' had shown that methyl cyanide combined with titanium and tin tetrachlorides, to form derivatives of the type SnCl,.2CH,CN. Anhydrous acetonitrile, made by the distillation of acetamide and phosphorus pentasulphide, and further purification of the product, was added to silicon tetrachloride chilled in ice; no action was apparent or heat generated. The precaution of chilling in ice was taken, because in the case of tin derivatives the amount of heat produced in the reaction was sufficient to decompose the products. Here, however, even when the ordinary temperature was resumed, no combination was effected. The slightest trace of moisture, however, serves 1 Ann. Chem. (Liebig), 73, 226.

2 Ann. Chem. (Liebig), 74, 86.

3 Thesis, University of Pennsylvania, 1896. 4 Ann. Chem. (Liebig), 106, 281.