Technique and condition
This smooth, white paper has been quite extensively worked, perhaps with a rag wrapped around the finger as well as with a brush. The paper was very wet when the earlier washes were applied: warm pink tones for the mountains and cooler blue for the sky. The browner pink washes which give more form to the misty mountains were applied to drier paper, while the blue ones that create a sense of depth in the lower mountains on the right were applied after the paper had been allowed to dry out. The sky and the water were worked on after the mountains in the middle of the sheet, both probably at the same time. The area of purplish pink wash on the left side, made from a red lake mixed with ultramarine, and the very hazy sun, are the areas which seems to have been worked with a rag rather than a brush.
At least one red lake was used here, as well as vermilion applied very sparingly for the pale pink areas. Indian yellow can be recognised by its appearance in ultraviolet light. Prussian blue was used for the greener blue sky and water. The red lake mixed with ultramarine was used to create the purplish pink colour in the water and applied at a late stage to the mountains. Yellow ochre was mixed with the red and pink pigments for the brown shore.
The identifications of the vermilion, Prussian blue and yellow ochre were in fact confirmed by removed tiny samples the size of a pin-point, and placing them in the sample chamber of a scanning electron microscope, under an X-ray beam. This beam interacts with the elements that make up each pigment, and the resulting spectrum makes it possible to work out which elements are present. Since it is already known that the washes are pure colours, it is then possible to work out exactly which pigment was used in each case. Visual identifications of these materials can then be made on other watercolours, when it is already known from examination at moderate magnification that the wash consists of a pure pigment and not a mixture. In a complex and finished watercolour with multiple overlying washes, it would be foolish to attempt such visual identification.