Any magnesia present in the filtrate of the oxalate of lime is by the addition of a solution of phosphate of soda separated as phosphate of ammonia and magnesia, after having stood twenty-four hours. The precipitate is filtered off, washed with water to which a little chloride of ammonia is added, dried, and after calcining the fiber and adding the filter ash, glow heated in the crucible. The glowed substance is weighed after cooling, and is pyrophosphate of magnesia, from which the magnesia or carbonate of magnesia is calculated stoichiometrically. All the ascertained sums must be multiplied by 2, if they are to correspond to the analyzed and weighed quantity of ash.

The second half of the filtrate is used for determining the small quantity of sulphate of lime still contained in the hydrochlorate solution. By adding chloride of barium solution the sulphuric acid is bound to the barytes and sulphate of baryta separates as white precipitate. This is separated by filtering, washed, dried and weighed in the customary manner. From the weight of the sulphate of baryta is then computed the weight of sulphate of lime, which has passed over into solution. The ascertained sum is also to be multiplied with 2.

The manufacture of roll tar paper from the roll paper was at first found to be difficult, as it was impossible to submerge a surface larger than from ten to fifteen square yards, rolled up, in the tar, because more would have required too large a pan. Besides this, the paper tears easily, when it is in the hot tar. All kinds of experiments were tried, in order to impregnate the surface of the paper without employing too large a pan.

The following method was tried at first: The roll paper was cut into lengths of ten yards, which were rolled up loosely, so that a certain space was left between the different coils. These loose rolls, of course, occupied much space and could be put into the tar only in a standing position, because in a horizontal one the several coils would have pressed together again. The loose roll was therefore slipped over a vertical iron rod fastened into a circular perforated wooden foot. The upper end of this iron rod ended in a ring, in which the hook of a chain or rope could be fastened. With the aid of a windlass the roll was raised or lowered. When placed in the pan with boiling tar, it was left there until thoroughly saturated. It was then taken out, placed upon a table, and the excess of tar allowed to drip off into a vessel underneath. After partially drying, the roll was spread out in open air, occasionally turned, until sufficiently dried, when it was rolled up again.

In order to neutralize the smeary, sticky condition of the surface and avoid the disagreeable drying in open air, the experiment of strewing sand on the sticky places was tried next. The weight of the paper was largely increased by the sand, and appeared considerably thicker. For this reason the method of sanding the paper was at once universally adopted. To dispense with the process of permitting the surplus tar to drip off, means were devised by which it was taken off by scrapers, or by pressing through rollers. The scrapers, two sharp edged rods fastened across the pan, were then so placed that the paper was drawn through them. The excess of tar adhering to its surface was thereby scraped off and ran back into the pan.

This work, however, was performed better and to more satisfaction by a pair of rollers fastened to the pan. These performed a double duty; thoroughly removed the tar from the surface and by reason of their pressure they caused a more perfect incorporation of the tar with the fibers of the paper. Finally, different factories employed different methods of manufacture, one of which was to cut the rolls into definite lengths of about ten yards; these were then rerolled very loosely and immersed in the hot tar until sufficiently saturated. The paper was then passed through the roller, much pressure exerted, and then loosely rolled up again. Being tarred once, it was then laid into a second pan with hot tar, reeled out after a time, strewn with sand, and rolled up again. Another method was to cut clothes lines into lengths of about fifteen yards, and at a distance of two inches have knots tied in them. The paper was cut in lengths of ten or fifteen yards, three pieces of the knotted clothes line were then rolled between the loose coils of paper, which was then submerged in the tar, which on account of the knots could penetrate the paper. The paper was next sanded by permitting its lower surface to pass over dry sand in a box standing on the floor.

A workman rolled off the paper, and with his hand he strews sand on the upper surface. The rolling taking place on the edge of a table, by means of a crank, the excess of sand dropped off.

It is said by this method two workmen, one of which tends to the rolling and sanding, the other turning the crank, could turn out eighty rolls per day. This method is still in use. It is useless to describe the many antiquated methods in vogue in smaller factories, and it can truthfully be said that nearly all of them are out of date. It appears to be the fact of almost all inventions that when reduced to practical use, the arrangements, apparatus, and working methods employed are generally of the most complicated nature, and time and experience only will simplify them. This has been also the case with the methods in the roofing paper industry, which are at present gradually being reduced to a practical basis. The method gradually adopted has been described in the preceding. The pan is of a certain length, whereby it becomes possible to saturate the paper by slowly drawing it through the heated tar. This is the chief feature. The work is much simplified thereby and the workmen need not dip their hands into the tar or soil them with it. The work of impregnating has become much cleaner and easier, while at the same time the tar can be heated to a much higher temperature.