實驗一研究（1）以詞為單位排列句子（一個詞與下一個詞之間有一個字的空間間隔）對讀出整個句子的認識閾有何影響，（2）雖有相同字數，但由5個詞（多單位）或3個詞（少單位）組成的句子有無認識閾的差別。在整個句子所佔的空間寬度加以控制的條件之下，實驗二研究上述兩變項以外還探討詞單位與詞單位之間的空間大小的影響。實驗三則對每一個字之大小加以控制並試圖減少極限法的一般缺點而研究實驗一的兩個變項，綜合三個實驗的結果，有下述傾向：若以詞單位安排句子則提高句子的認試閾。這結果究竟反映記憶的歷程就習慣的因素，仍待以後的研究。

　　　　　Experiment I investigated (1) the effect of aggregating Chinese characters constituting a word (while providing a space between words) on the recognition threshold of sentences and (2) the effect of varying the number of words within a sentence (with the number of characters kept constant) on the recognition threshold of sentences. In Experiment II the effect of manipulating the space between words as well as the effects of the variables of Experiment I on the recognition threshold of sentences were studied, while the span of each sentence was kept the same. The two variables of Experiment I were studied in Experiment III by controlling the size of Chinese characters and by minimizing artifacts inherent in the general limit method as a threshold measuring technique. It was found that, contrary to expectation, the effect of arranging Chinese words as units within a sentence was to raise the recognition threshold. Whether this finding reflects a memory process as affecting recognition' threshold or subject's established habit of reading awaits further experimentation.