Raring pupils' handwriting samples for legibility at fairly frequent intervals can yield information useful in motivating pupils and in establishing a basis for corrective teaching. Yet, in practice the awkwardness of using the typical legibility rating scales probably prevents most teachers from making such assessments very often. notion underlying this study was that once teachers have established a set of criteria for making legibility iudgments by using scales, they should be able to make reasonably reliable judgments even when the formal use of scales is discontinued. Handwriting samples were obtained by having 240 fourth and sixth grade pupils write a paragraph containing two sentences which include all the letters in the alphabet and which are the standard sentences for the California and Wisconsin scales, respectively: The quick brown fox just came over to greet the lazy poodle and The quick brown fox jumps over the lazy dog. Thus, there were two standard sentence samples from each pupil. Three judges independently rated the samples for legibility. First, the California scale, a 7-point scale that is provided as a part of the California Achievement Test battery (Tiegs & Clark, 1957), and the Wisconsin scale, a 7-point scale devised for the present smdy by procedures outlined by Herrick and Erlebacher (cf. Herrick, 1961, pp. 207-231), were used to assign a legibility rating to the appropriate sample from each pupil. Then the judges assigned a legibility rating of 1 to 7 to each sample without using a scale. Inter-judge reliability coefficients (Pearsonian) were .77, .72, and .69 with the California scale and 3 5 , 80. and .74 with the Wisconsin scale; the between-scales correlation was . 79 When raclnys were made without scales, the inter-judge coefficients were .79, .91, and .85 with the Cal~fornia sentence samples and .72, .74, and .86 for the Wisconsin sentence samples; the between-samples r was .82. judges did about as well, then, without the scales as with them, and there was no loss in the reliability of the ratings for individuals when the scales were not used. When each judge's ratings of the California sample sentences with and without the scale were correlated, the rs were .77, .83, and .76; with the Wisconsin sentences rs were .76, .85 and .73. These results indicate there was no significant loss in the magnitude of judges' reliability ratings when the use of scales was discontinued. implication appears to be that, given a background of experience in making such judgments, teachers do nor need to use scales to make reliable judgments regarding the legibiliy of pupils' handwriting. Further research is needed to determine optimum procedures for establishing criteria for judging. Meanwhile, a preliminary period of scale-based judging appears to be useful.