An examination of the internal structures (fig. 26) reveals two distinct types of poison glands, the acid-secreting and the alkaline-secreting glands, and a prominent poison reservoir. In addition, there is a small pair of accessory structures which have been called lubricating glands, on account of the supposed function of their product. The acid-secreting gland empties into the distal end of the poison reservoir which in turn pours the secretion into the muscular bulb-like enlargement at the base of the shaft. The alkaline secreting gland empties into the bulb ventrad of the narrow neck of the reservoir. The poison is usually referred to as formic acid. That it is not so easily explained has been repeatedly shown and is evidenced by the presence of the two types of glands. Carlet maintains that the product of either gland is in itself innocent,--it is only when they are combined that the toxic properties appear. The most detailed study of the poison of the honey-bee is that of Josef Langer (1897), who in the course of his work used some 25,000 bees. Various methods of obtaining the active poison for experimental purposes were used. For obtaining the pure secretion, bees were held in the fingers and compressed until the sting was exserted, when a clear drop of the poison was visible at its tip. This was then taken up in a capillary tube or dilute solutions obtained by dipping the tip of the sting into a definite amount of distilled water. An aqueous solution of the poison was more readily obtained by pulling out the sting and poison sacs by means of forceps, and grinding them up in water. The somewhat clouded fluid was then filtered one or more times. For obtaining still greater quantities, advantage was taken of the fact that while alcohol coagulates the poison, the active principle remains soluble in water. Hence the stings with the annexed glands where collected in 96 per cent alcohol, after filtering off of the alcohol were dried at 40° C., then rubbed to a fine powder and this was repeatedly extracted with water. Through filtering of this aqueous extract there was obtained a yellowish-brown fluid which produced the typical reactions, according to concentration of the poison. The freshly expelled drop of poison is limpid, of distinct acid reaction, tastes bitter and has a delicate aromatic odor. On evaporation, it leaves a sticky residue, which at 100 degrees becomes fissured, and suggests dried gum arabic. The poison is readily soluble in water and possesses a specific gravity of 1.1313. On drying at room temperature, it leaves a residue of 30 per cent, which has not lost in poisonous action or in solubility. In spite of extended experiments, Langer was unable to determine the nature of the active principle. He showed that it was not, as had been supposed, an albuminous body, but rather an organic base.