It has been already remarked that there is no agreement as to the composition of the colouring principle of ultramarine; some maintain that it is an iron compound, whilst others regard the iron found in ultramarine as an accidental impurity, which has no connection with the colour. Recently this point appears to have been decided from the results of many most accurate analyses of ultramarine. It has been found that its chemical composition shows the greatest similarity with the mineral nepheline. Nepheline is a double silicate, its composition is expressed by the following formula: Na₂O.SiO₂ + Al₂O₃.2SiO₂. By a comparison of the analyses of green and blue ultramarine with that of nepheline, it is seen that blue ultramarine may be regarded as nepheline combined with sodium pentasulphide, whilst green ultramarine is nepheline combined with sodium bisulphide. If this is the correct view, the two species of ultramarine have the following formulæ:— 2{Na₂O.SiO₂ + Al₂O₃.2SiO₂} + Na₂S₅. 2{Na₂O.SiO₂ + Al₂O₃.2SiO₂} + Na₂S₂. Ultramarine is one of the most permanent pigments. It is not altered by the substances with which a pigment generally comes in contact. Being a sulphur compound, it retains its colour completely in air containing sulphuretted hydrogen. By acids, even weak organic acids such as malic, citric or tartaric, it is rapidly decomposed, sulphuretted hydrogen being evolved, and a greyish white residue left. It is sometimes found that the sugar used in making lemonade produces a perceptible smell of sulphuretted hydrogen. This is due to ultramarine, which has been added to the sugar to hide its yellow colour. From the preceding account of the manufacture of ultramarine it will be understood that there exist only two species of ultramarine, green and blue. In commerce a large number of different shades are found, which are not to be regarded as pure ultramarine; they owe their particular shade to additions. By mixing in a white pigment, such as barytes or starch, the paler shades are obtained; by adding a small quantity of a pure red pigment a colour with a violet tinge is produced. In a similar manner any number of shades of ultramarine can be obtained. The following account is taken from a new comprehensive work by J. Wunder, on the different ultramarines.