1. INTRODUCTION The visual field surrounding us contains nemerous objects, elements, and spaces differing continuously in size between the smallest and largest sizes theoretically visible. From the undifferentiated mosaic of the visual field, one is compelled to select a figure on which attention concentrates and detects a visual object of a certain scale. At the same time, one perceives texture which is composed of smaller elements not perceived independently. The visual perception in the hierarchy of size-level may be schematically represented as in Fig. 1. In this figure, elements which are usually aggregated in the environment are represented by diagonal lines. Each line shows the relation between the observation distance (horizontal axis) and the size of projected image on the retina (vertical axis). The vertical axis is hypothetically divided into three perceptual categories: an array of in-dependent elements, texture and a smooth plane. Using this figure, we may know how each of the elements changes in appearance according to the observation distance or at a given distance, which elements appear as in-dependent elements, texture or smooth planes. If the boundaries between neighboring categories are defined, it should become a useful knowledge for environmental designers who wish to manipulate texture in the environment. The experiment reported below is a part of a first attempt to define these boundaries in a quantitative way. Although using abstract patterns, the present experiment provided a set of empirical understandings which may be interpreted for actual or simulated environment. 2. PSYCHOPHYSICAL EXPERIMENT USING DOT-PATTERNS The question of when we perceive texture (as opposed to perceiving independent elements or smooth planes) for aggregated elements in the environment was made operational by asking the same question for two dimensional dot-patterns. This experiment attempted to define the necessary conditions for us to sense texture visually. 2.1. Stimuli and Procedure Stimuli were sheets of mat surfaced "dot-patterns" arranged in many black or gray geometric figures on a white or gray background. The dot-patterns were systematically varied based on the following tow hypotheses I (1) The basic requirement for an optical pattern to be seen as texture is the awareness of the repetitive spatial variation in brightness over an area of a certain broadness. To sense the variation in brightness, the size of dark or bright dots and spacing between neighboring dots must be large enough, and the brightness contrast must be strong enough. The requirement of repetitiveness implies the necessary broadness of the area where sufficient repetitive variation can occur. Element size, interval between elements, brightness contrast and frame size were, therefore, taken as relevant variables. (2) Neighboring elements were required to unite together to form a sense of texture. The degree of unity between elements was regarded to be influenced by arrangement of elements and element shape as well as those variables already taken. The dot-patterns consisted of six types, totaling two hundred thirty-one variations (see Fig. 2 to Fig. 7, and Table 1). Using the experimental apparatus shown in Fig. 8, each of thirty observers was presented various dot-patterns and was asked to make three judgments i (1) whether or not the stimulus appears as a smooth plane of homogeneous gray color? (2) whether or not is appears as an array of independent elements? (3) whether or not it appears as texture? The order of presentation of dot-patterns was randomized for each subject. 2. 2. Results and Discussion Fig. 9 shows a part of the experimental results using the standard dot-patterns. In the figure, the proportion of positive responses to three questions were plotted against element size. In the graph, the judgment of "smooth plane" and "texture" as well as that of

(View PDF for the rest of the abstract.)