=7, without however giving the details of the experiments on which this number is based. On the other hand, Troost, in a paper upon the general history of lithia and its salts,* has objected to the method by which my determination of the equivalent was made, and has returned to a number near that originally given by Berzelius. Troost states that chlorid of lithium on being heated in the air loses chlorine and takes up oxygen, so that it must give by the method of Pelouze an atomic weight for the metal higher than the truth. This fact was distinctly noticed in my former paper, and it was stated that the decomposition might be prevented by addition of a little pure sal-ammoniac to the chlorid of lithium before heating. Troost objects to this, not that he has proved the method of correction defective, but that we cannot in the end tell whether the salt contains its full proportion of chlorine or not, unless the true equivalent of lithium-the constant we are in search of-be known. But it is to be remarked that the product of the exchange of chlorine for oxygen is caustic lithia, exhibiting a strong alkaline reaction. I have twice or thrice prepared chlorid of lithium, adding sal-ammoniac, and igniting in a well closed platinum crucible, and have always found that several grams dissolved in a very small quantity of water (the salt is extremely soluble) gave not the slightest alkaline or acid reaction with the most delicate vegetable colors. Troost himself adopts crystallized carbonate of lithia as the salt to be analyzed in order to determine the equivalent. He precipitates the carbonate, washes it thoroughly, diffuses it in water through which carbonic acid gas is passed until the salt dissolves, evaporates the solution until the carbonate is deposited as a crystalline powder, and dries this powder at 200°. He determines the lithia in one portion of the salt by evaporation with pure sulphuric acid, and the carbonic acid in another portion by noting the loss of weight on fusion with silicic acid. In this way he arrives at the number 6·6 (=82·5). No proof is offered that exposure to a temperature of 200° is capable of removing every trace of water and all carbonic acid over a single equivalent; yet, unless this be effected, the atomic weight of lithium will be brought out less than the truth. The same result will follow from the mechanical loss of the least drop of fluid during the effervescence of the carbonate with sulphuric acid or the subsequent evaporation of the sulphate of lithia; and, without feeling the slightest doubt of the manipulative skill of the French chemist, we must admit that, in so delicate a process as the determination of an atomic weight, the solution of a carbonate and evaporation of the solution-steps which are generally *Ann. de Chim. et de Phys., [3], t. LI, p. 108. looked upon as undesirable in the common course of analysisshould, if possible, be avoided. I have recently made a new determination of the equivalent, deriving it now, from experiments upon the sulphate of lithia; applying, however, a method avoiding as I hope the source of error to which Marignac has drawn attention; an error which threw much difficulty in the way of his successful estimation of the atomic weights of cerium, lanthanum, and didymium. If we add a salt of baryta in excess to a solution of any sulphate, the precipitate usually contains a small amount of the soluble barytic salt, which cannot be washed out, and which therefore increases the apparent amount of sulphuric acid present, if the latter be calculated from the weight of the sulphate of baryta, supposed pure. On the other hand, if the soluble sulphate be in excess, it will mix with the precipitate to some extent, and thus the proportion of sulphuric acid may be brought out higher or lower than the truth, as the equivalent of the base under examination is lower or higher than that of baryta. So that, if we wish to determine the atomic weight of lithium, as Berzelius did, by mixing the solution of a known amount of sulphate of lithia with chlorid of barium and weighing the sulphate of baryta precipitated, we are not certain that the weight of the latter really corresponds to the quantity of sulphuric acid in the salt analyzed. The same objection applies to Marignac's analysis of the sulphates of cerium and the allied metals. He there noted the volume of a solution of chlorid of barium of known strength required to precipitate a weighed portion of the sulphate; when a precipitate ceases to form, more or less chlorid of barium may have been used than is really equiv alent to the sulphuric acid present. The amount of the above error, must however be constant if the sulphate precipitated, the salt of baryta used, and the circumstances of precipitation be all the same. If the same salt of baryta be used to precipitate different sulphates, it is probable that the amount of error will be different for each. But, if we take the sulphates of two very similar and closely related bases, it is probable that the difference in the amount of error will be very small. These considerations have led to the following method for determining the equivalent of lithium. Sulphate of lithia was prepared, with all possible care, from the carbonate, and tested rigidly as to its purity. Two separate portions (A, 1, and 2,) of this salt were rendered anhydrous by cautious application of a heat below redness, and accurately weighed. Two similar portions of perfectly pure sulphate of soda (B, 1, and 2,) were dried and weighed with equal care. And, lastly, two portions of pure sulphate of magnesia (C, 1, and 2,) were in like manner dried and weighed. Soda and magnesia were chosen for comparison with lithia because the last-named base seems in most of its relations to hold an intermediate place between the former two, with which it is closely allied. Chlorid of barium was also prepared with all the precautions needed to ensure its purity, precipitated twice from its aqueous solution by alcohol, and recrystallized three or four times. It was at last obtained as a fine crystalline powder by stirring the hot saturated solution as it cooled, and this powder was allowed to dry spontaneously in the air at a temperature of about 80° F. Thus prepared, the salt-as Marignac has shown-is not altered in weight by further exposure to air, its theoretical composition is BaCl+2HO, the precise amount of water actually present was probably a little greater, owing to the mode of drying, but was unimportant under the conditions of experiment adopted. For each of the six weighed portions of sulphates mentioned above, the quantity of chlorid of barium needed for exact precipitation was calculated, assuming the equivalent of sodium =23, that of magnesium =12, that of lithium =7, and that of barium 686, and considering the chlorid of barium as containing strictly two atoms of water. Six portions of the lastnamed salt were weighed out (at the same time), each less than the amount calculated by one or two centigrams. Each was dissolved in 200 cubic centimetres of hot water, and added to its corresponding portion of sulphate, likewise dissolved in 200 cub. centim. of hot water. The fluid and precipitate in the six beakers were well stirred, and left to settle. = A solution was now prepared of 1 gram of the crystallized chlorid of barium (weighed out at the same time with the larger portions) in 1 litre of water, each cubic centimeter corresponding therefore to 1 miligram of BaCl+2HO. With this standard solution, dropped from a pipette whose degrees th of a cubic centim., the precipitation of the fluid in each of the six beakers was completed-the amount of chlorid of barium thus employed was noted, and added to the weight of the main portion originally taken. At first it was easy to observe the formation of a precipitate on each successive addition of the chlorid of barium solution, and subsidence took place quickly; but, as the point of exact neutralization was more and more nearly approached, each observation became more difficult, and hours and even days were required in order to observe the production of a cloud by each drop added, or to get the fluid clear again for another trial. When the last addition of chlorid of barium altogether failed to produce a precipitate, a single drop of a solution of sulphate of soda was added, and the formation of a cloud noticed. In this way the following results were obtained : : A, 1.-3.8924 grm. of LiO, SO, required for complete precipitation 8.6323 grm. of BaCl+2HO as used. A, 2.-4.6440 grm. of LiO, SO, required 10-2940 grm. of BaCl+2HO. B, 1.-5.0675 grm. of NaO, SO, required 8-6920 grm. of BaCl+2HO. B, 2.-5-1107 grm. of NaO, SO, required 8-7688 grm. of BaCl+2HO. C, 1.-4.3380 grm. of MgO, SO, required 8.8318 grm. of BaCl+2HO. C, 2.-4.6625 grm. of MgO, SO, required 9-4872 grm. of BaCl+2HO. Calculating now from B, and C, the amount of crystalline chlorid of barium necessary to precipitate an equivalent of NaO, SO, or MgO, SO,, we get the following numbers, which represent what may be called the practical equivalent of the chlorid of barium as actually used. the theoretical equivalent of BaCl+2HO being 122.1-the presence of any water over the normal two atoms tends to raise the practical equivalent-the presence of any BaCl in the precipitated BaO, SO, has the same effect, the presence of either of the soluble sulphates in the same precipitate leads to an opposite result. From this practical equivalent of chlorid of barium and the results given above under A, 1, and A, 2, we may calculate the equivalent of lithium. If we adopt for chlorid of barium the number 121 80-that obtained by the precipitation of NaO, SO3 -we have for A, 1, 54.95-48-6·95=Li. The mean of the two results is 6.935. If we take for chlorid of barium the number 122-12-derived from the experiments with MgO, SO,-we get by a similar calculation, Lastly, if we take the mean of the two numbers for chlorid of barium, namely, 121.96, we get for A, 1, or, in the mean, 7·005. Li=6.99 Hence, we find, that the equivalent of lithium, as deduced from the mean results of the above experiments, comes out 6.935 (86-69 on the oxygen scale) or as we take the practical equivalent, or actual precipitating power, of chlorid of barium from the experiments with NaO, SO,, those with MgO, SO,, or the mean of the two, these numbers exhibiting close agreement, and obviously indicate 7 as the true equivalent of the metal. It will be observed that the above method is independent of a knowledge of the exact equivalent of barium, and uses chlorid of barium merely as a means of bringing sulphate of lithia into comparison with the sulphates of soda and magnesia-the equivalents of the two last named bases may be considered as ranking among those best established -and the small difference between the practical equivalents for chlorid of barium deduced from these two shows the probable extent of error involved in the assumption of the same constant in the precipitate of the sulphate of lithia. While these results confirm those formerly obtained by the analysis of chlorid of lithium, I do not consider them of superior or perhaps even of equal value. The estimation of chlorine by the method of Pelouze is apparently one of the most simple and exact processes for the determination of an atomic weight which have ever been proposed, and it is, as I believe, fully applicable to the case of chlorid of lithium. As the result of both sets of experiments we may fairly take the number 7 (=87·50) as representing the true equivalent of the metal. ART. XXXVIII.-Notes on certain Ancient and Present Changes along the Coast of South Carolina; by OSCAR M. LIEBER, State Geologist, S. C. IT is very evident that remarkable changes have taken place on the coast of South Carolina during the present geological epoch; changes, which have effected or are yet, effecting very conspicuous alterations in the contour of the coast as well as in the hydrography of the immediate interior, and the elevation |