低照度下における横組みのローマ字および「かな」の活字の可読閾を測定して結果, 次の諸点が明らかにされた。(1)1音節を表わす場合と単語を表わす場合の何れにおいてもローマ字大文字の可読閾は小文字や「かな」の可読閾に比較して低い。(2)ローマ字小文字は母音を表わす場合は「かな」より可読閾が高いが, 母音以外の音節を表わす場合は「かな」よりも低い。また, 有意味単語を表わす場合は「ひらがな」よりは高いが「かたかな」と同程度の可読閾が得られ, 無意味音節を表わす場合は「ひらがな」と同程度で「かたかな」よりは低い可読閾が得られる。(3)単独の文字として呈示された場合は「ひらがな」も「かたかな」も可読閾においてほぼ等しい。単語して呈示される場合は前者の可読閾が方が低い。しかし, 外来語系統の単語では両者の差がない。(4)単独の文字として呈示された場合, 濁点・半濁点の存在は「かな」の可読閾を著しく高くする。特に「バ」行と「パ」行の「かな」の識別が困難である。しかし有意味単語においてはその影響が殆んどみられない。一方無意味音節においては単独に呈示された文字の場合と同様な影響がみられる。(5)ローマ字小文字は単独に呈示される場合よりも2字以上組み合わせて呈示した場合の方が可読閾が低い。この場合を除いては, ローマ字でも「かな」でも文字数と可読閾との間に一義的関係は認められない。(6)実験に用いられた5種類の音節に関する限り訓令式ローマ字と標準式ローマ字の可読閾に差はみとめられない。(7)単語の意味の有無によって「かな」の可読閾が異なるが, ローマ字の可読閾には殆んど影響がない。(8)ローマ字で表わされた国語の単語と英語の単語の可読閾に差は認められない。ただし, 固有名詞のようにローマ字の形で目に触れる機械の多い単語は一般的な国語の単語よりもローマ字の可読閾が低い。

In the present study the legibility of printed letters and words of Roma-ji and Japanese Kana were measured. Method : A 8.5×4.6 cm. white card on which black letter or letters were printed was presented to S in a visibility measuring apparatus. (see Fig. 1) This card was illuminated by a 100 watt Mazda lamp. The voltage of the lamp was varied from 0 to 70 v. by means of a rheostat in the circuit. The relationship between voltage and intensity of illumination was shown in Fig. 3. The exposure time of 10 sec. for one card was obtained by handling a sliding shutter. Sixteed-point type of Old style was used as stimulus in all experiments except Experiment 6, in which 14-point type was used. After 5 minutes of dark adaptation about 20 cards were presented one by one in random order under dim illumination, and S was asked to read the letter. After all cards were presented to S, the intensity of illumination was increased by 1 v. and cards were presented as before. In this way the minimum light intensity necessary for reading letters was measured. Except for Experiment 6. the legibility of each letter was measured twice. The Ss were male undergraduate and graduste students at the University of Tokyo. Results : (1) The capital letter of Roma-ji is the most legible among Kata-kana, Hira-gana, and capital and lower case of Roma-ji through all experiments. (2) The single lower case of Roma-ji representing each vowel is less legible than the single letter of Kata-kana or that of Hira-gana. The values of Table 1 are the pooled means, in voltage, of legibility of letters for 3 Ss. (3) With the exception of vowels, the single letter of Kata-kana and that of Hira-gana are less legiblethan the lower cases of Roma-ji representing the same monosyllable. The values of Table 5 are pooled means for 3 Ss. (4) There is no relationship between legibility of a syllable and number of letters of which it is constituted. The values of Table 9 are pooled means for 4 Ss. (5) The legibility of letters in Hepburnian system of Roma-ji does not differ from that of Japanese system. The mean values for 4 Ss. are shown in Table 11.