(6) Apply Gabbet's stain to the under side of the cover-glass to remove excess of carbol-fuchsin, and then to the film side. Allow this to act for one-fourth to one-half minute. (7) Wash in water. (8) If, now, the thinner portions of the film are blue, proceed to the next step; if they are still red, repeat steps (6) and (7) until the red has disappeared. Too long application of Gabbet's stain will decolorize the tubercle bacilli. (9) Place the preparation between layers of filter-paper and dry by rubbing with the fingers, as one would in blotting ink. (10) Put a drop of Canada balsam upon a clean slide, place the cover-glass film side down upon it, and examine with a one-twelfth objective. Cedar oil or water may be used in place of balsam for temporary preparations. Smears on slides may be examined directly with an oil-immersion lens, no cover being necessary. _Carbol-fuchsin_ is prepared by mixing 10 c.c. of a saturated alcoholic solution of fuchsin with 90 c.c. of 5 per cent. aqueous solution of phenol. {34} _Gabbet's stain_ consists of methylene-blue, 2 gm.; 25 per cent. sulphuric acid, 100 c.c. Both stains can be purchased ready prepared. * * * * * In films stained by Gabbet's method tubercle bacilli, if present, will be seen as slender red rods upon a blue background of mucus and cells (Plate II, Fig. 2). They average 3 to 4 µ in length--about one-half the diameter of a red blood-corpuscle. Beginners must be warned against mistaking the edges of cells, or particles which have retained the red stain, for bacilli. The appearance of the bacilli is almost always typical, and if there seems room for doubt, the structure in question is probably not a tubercle bacillus. They may lie singly or in groups. They are very frequently bent and often have a beaded appearance. It is possible that the larger, beaded bacilli indicate a less active tuberculous process than do the smaller, uniformly stained ones. Sometimes they are present in great numbers--thousands in a field of the one-twelfth objective. Sometimes several cover-glasses must be examined to find a single bacillus. At times they are so few that none are found in stained smears, and special methods are required to detect them. The number may bear some relation to the severity of the disease, but this relation is by no means constant. The mucoid sputum from an incipient case sometimes contains great numbers, while sputum from large tuberculous cavities at times contains very few. Failure to find them is not conclusive, though their absence is much more significant when the sputum is purulent than when it is mucoid. [Illustration: PLATE II. Fig. 1.--Heart-failure cells in sputum, containing blood-pigment, from a case of cardiac congestion of the lungs (Jakob).