nearly all removed, when more of the chloride seems to dissolve. This treatment with alcohol is continued until, on evaporation of a small portion of the latest filtrate, only a small residue is found. The alcoholic solution of the perchlorate is then distilled from a large flask until the perchlorate begins to crystallize, when the heat is removed and the contents quickly emptied into an evaporating dish, the same liquid being used to wash out the remaining portions of the salt. When the distillation is terminated at the point indicated, the distillate will contain most of the alcohol employed, but in a somewhat stronger solution, so that it requires only diluting to 97 per cent to fit it for use in future preparations. The salt is then evaporated to dryness on the steam bath and subsequently treated with strong hydrochloric acid for the separation of the perchloric acid. One cubic centimeter of the acid prepared in this way, on evaporation gave a residue in one case of 0.0369 grms., and in another 0-0307 grm., completely soluble in 97 per cent alcohol, which was then ignited and the chlorine determined by silver from which the equivalent of perchloric acid in the form of salts was calculated as 0-0305 grm. By neutralizing the acid with sodium carbonate, evaporating, igniting in an atmosphere of carbon dioxide till decomposition was complete, collecting the oxygen over caustic potash, allowing it to act on hydriodic acid by intervention of nitric oxide, according to a process soon to be published, titrating the iodin liberated, with standard arsenic and calculating the equivalent of perchloric acid, after subtracting the amount of acid found in the form of salts, the amount of free acid per cubic centimeter proved to be 0.9831 grms. The whole process, even when the separation with alcohol is necessary, can not well require more than two days and during the greater part of that time the work proceeds without attention. In applying perchloric acid, thus prepared, to the determination of potassium according to the treatment suggested by Caspari* very satisfactory results were obtained. Briefly, the method is as follows: The substance, free from sulphuric acid, is evaporated to the expulsion of free hydrochloric acid, the residue stirred with 20 cm3 of hot water and then treated with perchloric acid in quantity not less than one and one-half times that required by the bases present, when it is evaporated with frequent stirring to a thick, syrup-like consistency, again dissolved in hot water and evaporated with continued stirring till all hydrochloric acid has been expelled and the fumes of per * loc. cit. chloric acid appear. Further loss of perchloric acid is to be compensated for by addition of more. The cold mass is then well stirred with about 20 cm3 of wash alcohol-97 per cent alcohol containing 0-2 per cent by weight of pure perchloric acid-with precautions against reducing the potassium perchlorate crystals to too fine a powder. After settling, the alcohol is decanted on the asbestos filter and the residue similarly treated with about the same amount of wash alcohol, settling and again decanting. The residual salt is then deprived of alcohol by gently heating, dissolved in 10 cm3 of hot water and a little perchloric acid, when it is evaporated once more with stirring, until fumes of perchloric acid rise. It is then washed with 1 cm3 of wash alcohol, transferred to the asbestos, preferably by a policeman to avoid excessive use of alcohol, and covered finally with pure alcohol: the whole wash process requiring about 50 to 70 cm3 of alcohol. It is then dried at about 130° C. and weighed. The substitution of a Gooch crucible for the truncated pipette employed by Caspari will be found advantageous; and asbestos capable of forming a close, compact felt should be selected, inasmuch as the perchlorate is in part unavoidably reduced, during the necessary stirring, to so fine a condition that it tends to run through the filter when under pressure. A special felt of an excellent quality of asbestos was prepared for the determinations given below and seemed to hold the finer particles of the perchlorate very satisfactorily. A number of determinations made of potassium unmixed with other bases or non-volatile acids are recorded in the following table: Considerable difficulty, however, was experienced in obtaining equally satisfactory determinations of potassium associated with sulphuric and phosphoric acids. As Caspari has pointed out, the sulphuric acid must be removed by precipitation as barium sulphate before the treatment with perchloric acid is attempted, and unless the precipitation is made in a strongly AM. JOUR. SCI.-THIRD SERIES, VOL. XLIX, No. 294.-JUNE, 1895. acid solution, some potassium some potassium is carried down with the barium. Phosphoric acid need not be previously removed; but to secure a nearly complete separation of this acid from the potassium, a considerable excess of perchloric acid should be left upon the potassium perchlorate before it is treated with the alcohol. When these conditions are carefully complied with, fairly good results may justly be expected. Below are given a number of the results obtained. In the last three experiments of the above table the amount of perchloric acid was about three times that required to unite with the bases present and the phosphoric acid subsequently found with the potassium was hardly enough to appreciably affect the weight, although its absolute removal was found impossible. The kindly direction and frequent advice of Professor F. A. Gooch, during the investigation, is gratefully acknowledged. *The residue showed phosphoric acid plainly when tested. ART. XXXVII.-On the Crystal Form of Borneol and Isoborneol; by WM. H. HOBBS. 10 18 IN a recent paper by Bertram and Walbaum* on an isomer of borneol (C,,H,O) which they have called isoborneol, Traube has described both this substance and borneol from a crystallographical standpoint. The borneol examined was obtained by reduction of camphor, had a melting point of 206°-207°, and was dextro-rotatory. The symmetry of both borneol and isoborneol as determined by Traube is hexagonal, the combination in each case being the basal pinacoid with the pyramid and prism. The chief differences between the two substances he finds to be the greater double refraction of isoborneol, and its positive optical character, borneol being optically negative. The axial ratio of borneol he determined to be exactly double that of isoborneol. 10 18 Three samples of the alcohol C, II,,O were given me for examination to determine whether they are borneol or isoborneol. They were prepared in the School of Pharmacy of the University of Wisconsin by Mr. Carl G. Hunkel, whose study of them will be published in the Pharmaceutische Rundschau. The samples were prepared, one from the oil of black spruce (Picea nigra) in which the alcohol is contained as acetic ester, a second from the oil of the fir balsam (Abies balsamia), and the third from the oil of turpentine in benzine. The crystals in all these samples are larger and more highly modified than those described by Traube, and their examination has brought out new facts concerning their crystallography and physical properties. The surest basis of comparison with the crystals described by Traube has been the degree of double refraction. The crystals obtained from Picea nigra and Abies balsamia in this respect correspond exactly with the borneol of Traube's study. The crystals in the sample obtained from turpentine, on the other hand, correspond with his isoborneol so far as the degree of double refraction is concerned, but they are always optically negative, in this respect agreeing with borneol. It is therefore not certain that this substance is identical with the isoborneol of Bertram and Walbaum, but it seems best from all the facts to refer to it for the present as isoborneol. All the samples examined have rhombohedral symmetry. This is clearly shown by the partial occurrence of pyramids, and in the case of the crystals from Picea nigra by the tri-symmetric character of pittings on the basal pinacoid. Of the nine pyramidal forms which have been made out on the two substances * Ueber Isoborneol, Journ. f. prakt. Chemie, vol. xlix (1894), pp. 1-19. no one occurs in both positive and negative dodecants on the same crystal. The habit of both substances is broadly tabular parallel to the basal pinacoid and the plates have generally a regular hexagonal outline. One variety of isoborneol is, however, observed whose crystals take the form of rhomboidal plates owing to the disappearance of all planes from two of the opposite vertical pairs of dodecants. Although these crystals are identical with the normal variety in regard to their optical properties, they nevertheless represent an entirely different crystal combination. Crystals from all the samples have their faces more or less rounded and the measurements are as a result subject to considerable variations, but they are, nevertheless sufficiently accurate for a determination of all the forms. It is very probable that the axial ratios of borneol and isoborneol are different, since the substances differ so much in their double refraction, but they are certainly nearly identical and the difference is within the limits of error in the reading of angles on the crystals examined. I have therefore used for both substances the axial ratio determined on crystals of borneol from Picea nigra. Borneol from Picea nigra. The crystals of this substance examined are thin, colorless, hexagonal plates having a diameter of-1m. and a thickness of 0.5-1. The larger plates have a wide peripheral zone which is occupied by cavities generally filled with mother liquor. The shape of these cavities is somewhat irregular, but they are oriented roughly parallel to the boundaries of the plate. Besides the basal pinacoid the prominent forms are a steep rhombohedron making nearly 83° with the base and a smaller rhombohedral face of opposite sign which makes nearly 73° with the same form. This latter form is undoubtedly the pyramid observed on the substance by Traube and it is therefore chosen for determining the axial ratio. The average of four measurements of the angle included between this face and the base (limits 71° 25' and 74° 6') is 72° 46′ and if considered the fundamental rhombohedron the axial ratio would be = 2.79 (2.83, Traube). It is, however, more convenient to consider this form 3R (3031), which makes the axial ratio = 0.93. The observed forms are c, oP (0001); s, 3R (3031); 2, −8R (8051); m, ∞ P (1010); t, R (2023); u, 4R (4041). Figure 1 represents a crystal of borneol. These forms have been determined by the following measurements: |