Hitherto in considering the growth of the three-dimensional imagined picture-space we have not explicitly mentioned the part played by colour nor the equally important effect of this growth in modifying the original colours of the first retinal impression. But not only may colour be the chief factor determining form, i.e. the three-dimensional organisation of space, but it is itself most vitally modified by form. Colours as signs, that is to say even at the most optical and least elaborated stage, have certain very marked spatial characters of their own. Red, for example, seems to advance towards the eye and to swell out of its boundaries, while blue seems to retreat and to withdraw into itself[51]. Degree of saturation may also give recession in obvious and in more recondite ways. Pure colours in the foreground and greyed colours in the background are a simple example. Similarly opposition of colours is one of the main means by which the stresses and strains of volumes may be suggested. These characters of colours, especially when they reinforce and co-operate with one another, may be made to play a very important part in determining the way in which the picture-space is constructed when we look at a picture. Equally important are the less direct effects upon our picture-space imagining of the emotional or organic responses which we make to different colours. Individuals vary greatly in the extent to which they notice and can reflectively distinguish these responses, and probably also in the degree to which they actually make different responses. To persons sensitive in this respect, the colours excite each a distinct, well-marked emotion (and attitude) capable of being clearly differentiated from others. The sad poverty and vagueness of the colour vocabulary, however, misleads many people with regard to these. Each of the ‘puces’, ‘mauves’, ‘magentas’ etc. has to cover numbers of distinguishable colours, often with strikingly different effects upon us. Thus people who are content to say that pink is their favourite colour, or that green always suits them, are either quite undiscriminating in their attitude towards colour or little attentive to the actual effects produced upon them. A similar obtuseness or insincerity is evidenced when it is maintained, as is often done, that pink and green do not go together. Some pinks and some greens do not, but some do, and the test of a colourist is just his ability to feel which are which. Few if any, in fact, of the colour relations with which the painter is concerned can be stated with the aid of such general terms—‘red’, ‘brown’, ‘yellow’, ‘grey’, ‘primrose’, etc.—as are at present available. Each of these stands for a number of different colours whose relations to a given colour will commonly be different.